Funding, Deals & Partnerships: BIOLOGICS & MEDICAL DEVICES; BioMed e-Series; Medicine and Life Sciences Scientific Journal – http://PharmaceuticalIntelligence.com
In this blog post, I want to highlight some of the most important stories related to machine learning and NLP that I came across in 2019.
SAS AI/ML Training
By Elvis Saravia, Affective Computing & NLP Researcher
Other Related Topics
Activation Atlases is a technique developed by researchers at Google and Open AI to better understand and visualize the interactions happening between neurons of a neural network.
“An activation atlas of the InceptionV1 vision classification network reveals many fully realized features, such as electronics, buildings, food, animal ears, plants, and watery backgrounds.” — source
This Colab notebook provides a great introduction on how to use Nucleus and TensorFlow for “DNA Sequencing Error Correction”. And here is a great detailed post on the use of deep learning architectures for exploring DNA.
Alexander Rush is a Harvard NLP researcher who wrote an important article about the issues with tensors and how some current libraries expose them. He also went on to talk about a proposal for tensors with named indices.
ML/NLP Tools and Datasets ⚙️
StanfordNLP released StanfordNLP 0.2.0 which is a Python library for natural language analysis. You can perform different types of linguistic analysis such as lemmatization and part of speech recognition on over 70 different languages.
GQA is a visual question answering dataset for enabling research related to visual reasoning.
exBERT is a visual interactive tool to explore the embeddings and attention of Transformer language models. You can find the paper here and the demo here.
Natural language processing is perhaps the most talked-about subfield of data science. It’s interesting, it’s promising, and it can transform the way we see technology today. Not just technology, but it can also transform the way we perceive human languages.
Natural language processing has been gaining too much attention and traction from both research and industry because it is a combination between human languages and technology. Ever since computers were first created, people have dreamt about creating computer programs that can comprehend human languages.
The advances in machine learning and artificial intelligence fields have driven the appearance and continuous interest in natural language processing. This interest will only grow bigger, especially now that we can see how natural language processing could make our lives easier. This is prominent by technologies such as Alexa, Siri, and automatic translators.
The truth is, natural language processing is the reason I got into data science. I was always fascinated by languages and how they evolve based on human experience and time. I wanted to know how we can teach computers to comprehend our languages, not just that, but how can we make them capable of using them to communicate and understand us.
In this article, I will go through the 6 fundamental techniques of natural language processing that you should know if you are serious about getting into the field.
NLP 101: What is Natural Language Processing?
How did NLP start?
towardsdatascience.com
Lemmatization and stemming
Stemming and lemmatization are probably the first two steps to build an NLP project — you often use one of the two. They represent the field’s core concepts and are often the first techniques you will implement on your journey to be an NLP master.
Often, beginners tend to confuse the two techniques. Although they have their similarities, they are quite different.
Stemming: Stemming is a collection of algorithms that work by clipping off the end of the beginning of the word to reach its infinitive form.These algorithms do that by considering the common prefixes and suffixes of the language being analyzed. Clipping off the words can lead to the correct infinitive form, but that’s not always the case. There are many algorithms to perform stemming; the most common one used in English is the Porter stemmer. This algorithm contains 5 phases that work sequentially to obtain the word’s root.
Lemmatization: To overcome the flaws of stemming, lemmatization algorithms were designed. In these types of algorithms, some linguistic and grammar knowledge needs to be fed to the algorithm to make better decisions when extracting a word’s infinitive form. For lemmatization algorithms to perform accurately, they need to extract the correct lemma of each word. So, they often require a dictionary of the language to be able to categorize each word correctly.
Based on these definitions, you can imagine that building a lemmatizer is more complex and more time consuming than building a stemmer. However, it is more accurate and will cause less noise in the final analysis results.
Keywords extraction
Keyword extraction — sometimes called keyword detection or keyword analysis — is an NLP technique used for text analysis. This technique’s main purpose is to automatically extract the most frequent words and expressions from the body of a text. It is often used as a first step to summarize the main ideas of a text and to deliver the key ideas presented in the text.
In the backend of keyword extraction algorithms lays the power of machine learning and artificial intelligence. They are used to extract and simplify a given text for it to be understandable by the computer. The algorithm can be adapted and applied to any type of context, from academic text to colloquial text used in social media posts.
Keywords extraction has many applications in today’s world, including social media monitoring, customer service/ feedback, product analysis, and search engine optimization.
Named Entity Recognition (NER)
Like stemming and lemmatization, named entity recognition, or NER, NLP’s basic and core techniques are. NER is a technique used to extract entities from a body of a text used to identify basic concepts within the text, such as people’s names, places, dates, etc.
NER algorithm has mainly two steps. First, it needs to detect an entity in the text and then categorize it into one set category. The performance of NER depends heavily on the training data used to develop the model. The more relevant the training data to the actual data, the more accurate the results will be.
another factor contributing to the accuracy of a NER model is the linguistic knowledge used when building the model. That being said, there are open NER platforms that are pre-trained and ready to use.
NER can be used in a varsity of fields, such as building recommendation systems, in health care to provide better service for patients, and in academia to help students get relevant materials to their study scopes.
Topic Modelling
You can use keyword extractions techniques to narrow down a large body of text to a handful of main keywords and ideas. From which, you can probably extract the main topic of the text.
Another, more advanced technique to identify a text’s topic is topic modeling—top modeling built upon unsupervised machine learning that doesn’t require a labeled data for training.
Multiple algorithms can be used to model a topic of text, such as Correlated Topic Model, Latent Dirichlet Allocation, and Latent Sentiment Analysis. The most commonly used approach is the Latent Dirichlet. This approach analyzes the text, breaks it down into words and statements, and then extracts different topics from these words and statements. All you need to do is feed the algorithm a body of text, and it will take it from there.
One of the useful and promising applications of NLP is text summarization. That is reducing a large body of text into a smaller chuck containing the text’s main message. This technique is often used in long news articles and to summarize research papers.
Text summarization is an advanced technique that used other techniques that we just mentioned to establish its goals, such as topic modeling and keyword extraction. The way this is established is via two steps, extract and then abstract.
In the extract phase, the algorithms create a summary by extracting the text’s important parts based on their frequency. After that, the algorithm generates another summary, this time by creating a whole new text that conveys the same message as the original text. There are many text summarization algorithms, e.g., LexRank and TextRank.
In LexRank, the algorithm categorizes the sentences in the text using a ranking model. The ranks are based on the similarity between the sentences; the more similar a sentence is to the rest of the text, the higher it will be ranked.
Sentiment Analysis
The most famous, well-known, and used NLP technique is, without a doubt, sentiment analysis. This technique’s core function is to extract the sentiment behind a body of text by analyzing the containing words.
The technique’s most simple results lay on a scale with 3 areas, negative, positive, and neutral. The algorithm can be more complex and advanced; however, the results will be numeric in this case. If the result is a negative number, then the sentiment behind the text has a negative tone to it, and if it is positive, then some positivity in the text.
Sentiment analysis is one of the broad applications of machine learning techniques. It can be implemented using either supervised or unsupervised techniques. Perhaps the most common supervised technique to perform sentiment analysis is using the Naive Bayes algorithm. Other supervised ML algorithms that can be used are gradient boosting and random forest.
Humans’ desire for computers to understand and communicate with them using spoken languages is an idea that is as old as computers themselves. Thanks to the rapid advances in technology and machine learning algorithms, this idea is no more just an idea. It is a reality that we can see and experience in our daily lives. This idea is the core diving power of natural language processing.
Natural language processing is one of today’s hot-topics and talent-attracting field. Companies and research institutes are in a race to create computer programs that fully understand and use human languages. Virtual agents and translators did improve rapidly since they first appeared in the 1960s.
Despite the different tasks that natural language processing can execute, to get in the field and start building your own projects, you need to be completely comfortable with the core 6 fundamental natural language processing techniques.
These techniques are the basic building blocks of most — if not all — natural language processing algorithms. So, if you understand these techniques and when to use them, nothing can stop you.
A Learning Path To Becoming a Data Scientist
The 10 steps roadmap to kickstarting your data science future
Google Cloud launches Vaccine Management Tools using ML & AI for Vaccine Distribution Efforts
Reporter: Aviva Lev-Ari, PhD, RN
Google Cloud announced Monday new artificial intelligence and machine learning tools to help with vaccine rollout efforts from vaccine information and scheduling, to distribution and analytics, to forecasting and modeling COVID-19 cases.
Google Cloud is the latest company to put its tech muscle behind efforts to vaccinate millions of Americans against the COVID-19 virus.
The company launched artificial intelligence and machine learning tools Monday to help organizations forecast and model COVID-19 cases to better inform vaccine allocation. The cloud-based tools also are designed to assist with vaccine distribution, appointment scheduling, eligibility screening and communications.
The technology, called the Intelligent Vaccine Impact solution, also analyzes consumer sentimentaround the COVID-19 vaccine. Understanding how local communities feel about the risks and benefits of the vaccine is critical to being able to increase confidence in vaccination, Google Cloud executives said. The sentiment analysis tool, which was developed in partnership with behavioral intelligence data company Syntasa, will help public health agencies develop a more tailored and informed vaccination campaign.
“We’ve created a set of core technologies to help regional and local governments deliver successful COVID-19 public health strategies, ranging from vaccine information and scheduling, to distribution and analytics, to forecasting and modeling COVID-19 cases,” said Mike Daniels, vice president of the global public sector at Google Cloud, in a blog post.
The new solution helps increase vaccine availability and equitable access to those who need it and assists governments in building awareness, confidence and acceptance of vaccines, Daniels wrote.
“We designed our solution to easily integrate with existing technologies, knowing that governments will administer their vaccine distributions in unique ways,” he said.
As the COVID-19 vaccine distribution effort in the U.S. gets off to a bumpy start, tech giants are seizing the business opportunity and throwing their considerable resources into the effort to overcome logistical hurdles. Microsoft is partnering with a number of organizations on vaccine management efforts for both government and healthcare customers. The tech giant worked with business partners including Accenture, Avanade, EY and Mazik Global to deploy vaccine management solutions that enable registration capabilities for patients and providers, phased scheduling for vaccinations, streamlined reporting and management dash-boarding with analytics and forecasting.
IBM launched supply chain management software and a blockchain-based approach to verify and track vaccines, certify that patients have had the vaccine and keep track of which patients received which vaccine.
Salesforce rolled out a cloud-based vaccine management solution that’s being used by government agencies and healthcare organizations including Northwell Health, Illinois’ Lake County, University of Massachusetts Amherst and Gavi, the Vaccine Alliance. And Oracle set up an electronic health record database and public health applications to track vaccinations and side effects.
As part of its efforts, Google Cloud researchers developed a machine learning approach that combines AI with a foundation of epidemiology to help with COVID-19 forecasting, Daniels said in the blog post. Researchers also developed an AI-driven “what-if” model to be used for COVID-19 response and other infectious disease policy intervention decision-making.
Government leaders can use data to see how forecasts change in response to policy changes, such as mask mandates, modified reopening plans or vaccination programs.
Google Cloud’s solution also includes a vaccine information portal to help overwhelmed local governments address questions and concerns from the public, Daniels said.
“Working in partnership with SpringML, MTX, Deloitte and other partners, Google Cloud has built several vaccine information portals that help people learn about vaccine availability, determine if they qualify, sign up for vaccination, and submit their information so that when they are eligible they can be vaccinated as quickly as possible,” he said in the blog post.
The solution also offers consumers online registration and prescreening, location searches and appointment setting as well as automated reminders. The solution relies on Google Cloud Healthcare API to transmit data using common formats such as HL7 or FHIR—which interoperate with existing healthcare and immunization systems, the tech giant said.
Several states including North Carolina and New York have already deployed Google Cloud’s vaccine management solution.
“Our newest effort is to develop a process and technology to streamline accessing information for North Carolinians,” said Sam Gibbs, deputy secretary for technology and operations for the state of North Carolina, in a statement. “This technology will provide a central location for residents to find information such as when it is their turn to get their vaccine or guidance to easily locate a vaccination location.”
Results of the survey appear in Figure 1 for the overall sample as well as countries that have 50 or more respondents. Overall, results show that the adoption rate of machine learning methods is 45%. Twenty-one percent of respondents indicate their company is exploring ML methods. Twenty percent of respondents indicate their company does not use ML methods.
Countries that are early adopters of ML methods include:
Israel (63% adopt ML)
Netherlands (57%)
United States (56%)
UK and Northern Ireland (54%)
Germany (54%)
Australia (53%)
France (52%)
China (52%)
Taiwan (51%)
Greece (49%)
Countries with the lowest adoption rate of ML methods include:
Nigeria (23% adopt ML)
Morocco (24%)
Egypt (31%)
Philippines (31%)
Argentina (32%)
Countries with the highest percent of companies exploring ML methods include:
A worldwide survey of data professionals showed that adoption of machine learning methods in their company is 45%. Twenty-one percent of survey respondents said their employer is exploring ML methods. ML adoption rates varied by country with Israel (63%), Netherlands (57%) and the United States (56%) showing the highest and Egypt (31%), Morocco (24%) and Nigeria (23%) showing the lowest adoption rate. ML adoption also varied by company size, with larger companies having higher adoption rates (61%) than medium (45%) and small (33%) companies.
Businesses are leveraging the power of machine learning methods to help them extract better quality information, increase productivity, reduce costs and extract more value from their data. As the amount of data continues to grow along with the processing power of technology, businesses will continue to incorporate ML into their business. Researchers have found different AI / ML adoption rates. In one study, adoption rate of ML Methods was 10%; in a 2020 study by McKinsey, adoption rate of AI was 50%. Still, another study found that 42% of companies were currently using AI and 40% of companies were planning on using AI in the next two years. Another 2020 study found that 59% of enterprises have machine learning initiatives either in production or at a proof-of-concept stage.
Current Analysis on Machine Learning Adoption
Kaggle conducted a worldwide survey in October 2020 of 20,036 data professionals (2020 Kaggle Machine Learning and Data Science Survey). The survey sample consisted of data professionals, including men (~79%) and women (~19%), from a variety of job titles (e.g., data scientist, business analyst, machine learning engineer, software developer) and company sizes. The survey asked a variety of questions, including “Does your current employer incorporate machine learning methods into their business?”
Figure 1. Machine Learning Adoption Rates across Countries. Click image to enlarge.
Figure 2. Adoption of ML Methods Across Company Size
We also looked at adoption rates by company size. Those results appear in Figure 2. Supporting prior studies, we found that larger companies have higher adoption rates about ML methods. The largest enterprise companies (10,000+ employees) reported ML adoption rates of 61%. The smallest companies (0-49 employees) reported adoption rates of 33%. Of the smallest companies, a little over a quarter of them (27%) indicate that they are exploring the use of ML methods.
Summary
Survey of data professionals showed that adoption rates of machine learning methods among businesses is 45%. About 21% of respondents indicated that their company is exploring machine learning methods with the hope of putting a model into production one day.
ML adoption rate varies by country and company size. Survey results reveal that early adopters come from large enterprise companies (adoption rate of 61%) and some countries including the United States, Israel, Netherlands and the UK and Northern Ireland.
Machine learning vendors, looking for inroads into businesses, could focus their marketing and sales efforts on small businesses as they have the highest percentage of companies who are exploring the use of ML methods.
Recent advances in biomedical informatics, such as the application of machine learning and natural language processing (NLP) to Electronic Health Record (EHR) data, have resulted in breakthroughs in the ability to identify, classify, and predict aspects of patient care, symptoms, and communication in the clinical record. These new tools that allow researchers and clinicians to generate and examine new research questions examining patient care quality and health-related quality of life.
This webinar presents four research studies highlighting different NLP applications in EHR free text. Topics include (1) phenotyping for treatment-related side effects and system severity, (2) cancer treatment-related side-effects, (3) non-adherence to cancer therapy and recurrence, and (4) documentation of patient care goals.
Presenters discuss how these methodologies support the capture and use of real-world data for precision oncology care, population health management, and how to scale-up applications within and across health care systems.
Presenters
Julian Hong, MD, MS
Assistant Professor
Healthcare Delivery Research Program
Department of Radiation Oncology
UCSF Medical Center
Charlotta Lindvall, MD, PhD
Instructor, Medicine, Harvard Medical School
Physician, Department of Psychosocial Oncology and Palliative Care
Dana-Farber Cancer Institute
Tina Hernandez-Boussard, PhD
Associate Professor
Stanford University School of Medicine
Imon Banerjee, PhD
Assistant Professor
Department of Biomedical Informatics
Emory University School of Medicine
Intended Audience
Extramural investigators at junior, mid, and senior levels. Basic familiarity with data science methods may be helpful, but not required.
re:Invent 2020 – Virtual 3 weeks Conference, Nov. 30 – Dec. 18, 2020: How Healthcare & Life Sciences leaders are using AWS to transform their businesses and innovate on behalf of their customers.
Preview the tracks that will be available, along with the general agenda, on the website.
AWS re:Invent routinely fills several Las Vegas venues with standing-room only crowds, but we are bringing it to you with an all-virtual and free event this year. This year’s conference is gearing up to be our biggest yet and we have an exciting program planned with five keynotes, 18 leadership sessions, and over 500 breakout sessions beginning November 30. Hear how AWS experts and inspiring members of the Life Sciences & Genomics industry are using cloud technology to transform their businesses and innovate on the behalf of their customers.For Life Sciences attendees looking to get the most out of their experience, follow these steps:
Take a look at all of the Life Sciences sessions available, as well as lots of other information and additional activities, in our curated Life Sciences Attendee Guide coming soon!
Check back on this post regularly, as we’ll continually update it to reflect the newest information.
Life Sciences at re:Invent 2020
AWS enables pharma and biotech companies to transform every stage of the pharma value chain, with services that enhance data liquidity, operational excellence, and customer engagement. AWS is the trusted technology provider with the cost-effective compute and storage, machine learning capabilities, and customer-centric know how to help companies embrace innovation and bring differentiated therapeutics to market faster.
Dates and Presentation Title
BlockChain Architecture
DEC 1, 2020 | 12:00 AM – 12:20 AM EST
Nestlé brings supply chain transparency with Amazon Managed Blockchain
Chain of origin is the Nestlé answer to complete supply chain transparency, from crop to coffee cup, with technology at its heart. Today, consumers want to know about the quality of their product and know where it is sourced from. Using AWS and Amazon Man
This session demonstrates how to build interactive, trusted, and transparent live virtual experiences using the Amazon Interactive Video Service, Amazon Chime, Amazon QLDB, and Amazon Managed Blockchain to capture cryptographic, immutable, and verifiable records of real-time audience interactions.
Add To Calendar
Architecture Blockchain
DEC 10, 2020 | 11:15 PM – 11:45 PM EST
Trust and transparency in live virtual events
This session demonstrates how to build interactive, trusted, and transparent live virtual experiences using the Amazon Interactive Video Service, Amazon Chime, Amazon QLDB, and Amazon Managed Blockchain to capture cryptographic, immutable, and verifiable records of real-time audience interactions.
Add To Calendar
Architecture Blockchain
Friday, December 11
DEC 11, 2020 | 7:15 AM – 7:45 AM EST
Trust and transparency in live virtual events
This session demonstrates how to build interactive, trusted, and transparent live virtual experiences using the Amazon Interactive Video Service, Amazon Chime, Amazon QLDB, and Amazon Managed Blockchain to capture cryptographic, immutable, and verifiable records of real-time audience interactions.
Add To Calendar
Architecture Blockchain
Tuesday, December 15
DEC 15, 2020 | 4:00 PM – 4:30 PM EST
What’s new in Amazon Managed Blockchain
Amazon Managed Blockchain is a fully managed service that makes it easy for you to create and manage scalable blockchain networks using popular open-source technologies. Blockchain technologies enable groups of organizations to securely transact and
Amazon Managed Blockchain is a fully managed service that makes it easy for you to create and manage scalable blockchain networks using popular open-source technologies. Blockchain technologies enable groups of organizations to securely transact and
Amazon Managed Blockchain is a fully managed service that makes it easy for you to create and manage scalable blockchain networks using popular open-source technologies. Blockchain technologies enable groups of organizations to securely transact and
Building robots to help hospitals become safer and smarter
In hospitals and other healthcare venues, robots increasingly perform contactless delivery and autonomous maintenance services to reduce the risk of exposing patients and medical staff to harmful viruses and bacteria. In this session, see how Solaris JetBrain and M
Transform research environments with Service Workbench on AWS
Reenvison how research environments are spun up by reducing wait times from days to minutes. Service Workbench on AWS promotes repeatability, multi-institutional collaboration, and transparency in the research process. In this session, learn how Harvard Medical School is procuring and deploying domain-specific data, tools, and secure IT environments to accelerate research.
Add To Calendar
Public Sector Healthcare
Wednesday, December 2
DEC 2, 2020 | 1:15 AM – 1:45 AM EST
Transform research environments with Service Workbench on AWS
Reenvison how research environments are spun up by reducing wait times from days to minutes. Service Workbench on AWS promotes repeatability, multi-institutional collaboration, and transparency in the research process. In this session, learn how Harvard Medical School is procuring and deploying domain-specific data, tools, and secure IT environments to accelerate research.
Add To Calendar
Public Sector Healthcare
DEC 2, 2020 | 9:15 AM – 9:45 AM EST
Transform research environments with Service Workbench on AWS
Reenvison how research environments are spun up by reducing wait times from days to minutes. Service Workbench on AWS promotes repeatability, multi-institutional collaboration, and transparency in the research process. In this session, learn how Harvard Medical School is procuring and deploying domain-specific data, tools, and secure IT environments to accelerate research.
Add To Calendar
Public Sector Healthcare
DEC 2, 2020 | 11:00 AM – 11:30 AM EST
Reinventing medical imaging with machine learning on AWS
It is hard to imagine the future of medical imaging without machine learning (ML) as its central innovation engine. Countless researchers, developers, startups, and larger enterprises are engaged in building, training, and deploying ML solutions for medical i
Making healthcare more personal with MetroPlus Health
COVID has made a huge impact across the world, and organizations have had to adapt quickly to changing requirements as a result. Learn how MetroPlus Health, a New York City health plan covering over half a million people, leveraged AWS technology to quickly
Securing protected health information and high-risk datasets
Join this session featuring Jonathan Cook, Chief Technology Officer at Arcadia, for a discussion around securing mission-critical and high-risk datasets such as personal health information (PHI) in the cloud. Learn how Arcadia developed a HITRUST CSF-certified plat
ML and analytics addressing nationwide COVID-19 impact and recovery (sponsored …
In this session, learn how Fractal.AI delivered a platform to help analyze data and make decisions related to COVID-19 progression for the government of Telangana, India, deploying it on AWS in five days. The solution, based on Intel processors, delivered more than 100 dashboards using anonymized government and public datasets with hundreds of thousands of COVID-19 d… Learn More
Add To Calendar
Public Sector Partner Solutions for Business
DEC 2, 2020 | 6:30 PM – 7:00 PM EST
How Vyaire uses AWS analytics to scale ventilator production & save lives
When COVID-19 hit the US, Vyaire Medical, one of the country’s only ventilator manufacturers, knew it would have to scale rapidly while still offering quality machinery. In order to scale to 20 times more than its usual production and gain insights from all its new
Reinventing medical imaging with machine learning on AWS
It is hard to imagine the future of medical imaging without machine learning (ML) as its central innovation engine. Countless researchers, developers, startups, and larger enterprises are engaged in building, training, and deploying ML solutions for medical i
Making healthcare more personal with MetroPlus Health
COVID has made a huge impact across the world, and organizations have had to adapt quickly to changing requirements as a result. Learn how MetroPlus Health, a New York City health plan covering over half a million people, leveraged AWS technology to quickly
Securing protected health information and high-risk datasets
Join this session featuring Jonathan Cook, Chief Technology Officer at Arcadia, for a discussion around securing mission-critical and high-risk datasets such as personal health information (PHI) in the cloud. Learn how Arcadia developed a HITRUST CSF-certified plat
ML and analytics addressing nationwide COVID-19 impact and recovery (sponsored …
In this session, learn how Fractal.AI delivered a platform to help analyze data and make decisions related to COVID-19 progression for the government of Telangana, India, deploying it on AWS in five days. The solution, based on Intel processors, delivered more than 100 dashboards using anonymized government and public datasets with hundreds of thousands of COVID-19 d… Learn More
Add To Calendar
Public Sector Partner Solutions for Business
DEC 3, 2020 | 2:30 AM – 3:00 AM EST
How Vyaire uses AWS analytics to scale ventilator production & save lives
When COVID-19 hit the US, Vyaire Medical, one of the country’s only ventilator manufacturers, knew it would have to scale rapidly while still offering quality machinery. In order to scale to 20 times more than its usual production and gain insights from all its new
Reinventing medical imaging with machine learning on AWS
It is hard to imagine the future of medical imaging without machine learning (ML) as its central innovation engine. Countless researchers, developers, startups, and larger enterprises are engaged in building, training, and deploying ML solutions for medical i
Making healthcare more personal with MetroPlus Health
COVID has made a huge impact across the world, and organizations have had to adapt quickly to changing requirements as a result. Learn how MetroPlus Health, a New York City health plan covering over half a million people, leveraged AWS technology to quickly
Securing protected health information and high-risk datasets
Join this session featuring Jonathan Cook, Chief Technology Officer at Arcadia, for a discussion around securing mission-critical and high-risk datasets such as personal health information (PHI) in the cloud. Learn how Arcadia developed a HITRUST CSF-certified plat
ML and analytics addressing nationwide COVID-19 impact and recovery (sponsored …
In this session, learn how Fractal.AI delivered a platform to help analyze data and make decisions related to COVID-19 progression for the government of Telangana, India, deploying it on AWS in five days. The solution, based on Intel processors, delivered more than 100 dashboards using anonymized government and public datasets with hundreds of thousands of COVID-19 d… Learn More
Add To Calendar
Public Sector Partner Solutions for Business
DEC 3, 2020 | 10:30 AM – 11:00 AM EST
How Vyaire uses AWS analytics to scale ventilator production & save lives
When COVID-19 hit the US, Vyaire Medical, one of the country’s only ventilator manufacturers, knew it would have to scale rapidly while still offering quality machinery. In order to scale to 20 times more than its usual production and gain insights from all its new
Productionizing R workloads using Amazon SageMaker, featuring Siemens
R language and its 16,000+ packages dedicated to statistics and ML are used by statisticians and data scientists in industries such as energy, healthcare, life science, and financial services. Using R, you can run simulations and ML securely and at scale with Amazon S
Accelerating the transition to telehealth with AWS
Learn how AWS is helping healthcare organizations develop and deploy telehealth solutions quickly and at scale. Join speakers from AWS and MedStar Health as they discuss their experience developing and deploying two call centers in less than a week using Amaz
Productionizing R workloads using Amazon SageMaker, featuring Siemens
R language and its 16,000+ packages dedicated to statistics and ML are used by statisticians and data scientists in industries such as energy, healthcare, life science, and financial services. Using R, you can run simulations and ML securely and at scale with Amazon S
Accelerating the transition to telehealth with AWS
Learn how AWS is helping healthcare organizations develop and deploy telehealth solutions quickly and at scale. Join speakers from AWS and MedStar Health as they discuss their experience developing and deploying two call centers in less than a week using Amaz
Productionizing R workloads using Amazon SageMaker, featuring Siemens
R language and its 16,000+ packages dedicated to statistics and ML are used by statisticians and data scientists in industries such as energy, healthcare, life science, and financial services. Using R, you can run simulations and ML securely and at scale with Amazon S
Accelerating the transition to telehealth with AWS
Learn how AWS is helping healthcare organizations develop and deploy telehealth solutions quickly and at scale. Join speakers from AWS and MedStar Health as they discuss their experience developing and deploying two call centers in less than a week using Amaz
How Erickson Living built a COVID-19 outbreak management solution
As innovators in independent living, assisted living, and skilled nursing care, Erickson Living responded to the COVID-19 outbreak by building an infectious disease management system with the AWS Data Lab to prevent the spread of the virus. This session focuses o
Rapidly deploying social services on Amazon Connect
Organizations that respond to disruptive, large-scale events need the ability to rapidly scale and iterate on their contact centers to provide services to their constituents. Amazon Connect can be set up in minutes and scale to handle virtually any number of contact
Improving data liquidity in Roche’s personalized healthcare platform
Roche’s personalized healthcare mission is to accelerate drug discovery and transform the patient journey by using digital technologies and advanced analytics to facilitate greater scientific collaboration and insight sharing. In this session, Roche shares ho
Life Sciences Artificial Intelligence & Machine Learning
DEC 8, 2020 | 10:30 PM – 11:00 PM EST
How Erickson Living built a COVID-19 outbreak management solution
As innovators in independent living, assisted living, and skilled nursing care, Erickson Living responded to the COVID-19 outbreak by building an infectious disease management system with the AWS Data Lab to prevent the spread of the virus. This session focuses o
Rapidly deploying social services on Amazon Connect
Organizations that respond to disruptive, large-scale events need the ability to rapidly scale and iterate on their contact centers to provide services to their constituents. Amazon Connect can be set up in minutes and scale to handle virtually any number of contact
Improving data liquidity in Roche’s personalized healthcare platform
Roche’s personalized healthcare mission is to accelerate drug discovery and transform the patient journey by using digital technologies and advanced analytics to facilitate greater scientific collaboration and insight sharing. In this session, Roche shares ho
Life Sciences Artificial Intelligence & Machine Learning
DEC 9, 2020 | 6:30 AM – 7:00 AM EST
How Erickson Living built a COVID-19 outbreak management solution
As innovators in independent living, assisted living, and skilled nursing care, Erickson Living responded to the COVID-19 outbreak by building an infectious disease management system with the AWS Data Lab to prevent the spread of the virus. This session focuses o
Rapidly deploying social services on Amazon Connect
Organizations that respond to disruptive, large-scale events need the ability to rapidly scale and iterate on their contact centers to provide services to their constituents. Amazon Connect can be set up in minutes and scale to handle virtually any number of contact
Improving data liquidity in Roche’s personalized healthcare platform
Roche’s personalized healthcare mission is to accelerate drug discovery and transform the patient journey by using digital technologies and advanced analytics to facilitate greater scientific collaboration and insight sharing. In this session, Roche shares ho
Life Sciences Artificial Intelligence & Machine Learning
DEC 9, 2020 | 10:30 AM – 11:00 AM EST
BlueJeans’ explosive growth journey with AWS during the pandemic
Global video provider BlueJeans (a Verizon company) supports employees working from home, healthcare providers shifting to telehealth, and educators moving to distance learning. With so many people now working from home, BlueJeans saw explosive gro
Learn how healthcare organizations can harness the power of AI and machine learning to automate clinical workflows, digitize medical information, extract and summarize medical information, protect patient data, and more. Using AWS document understand
As clinical trials increasingly become decentralized and virtual, engaging effectively with patients can be challenging. In this session, hear how Evidation Health architects on AWS to create patient-centric experiences, ingests data from millions of devices in real time
BlueJeans’ explosive growth journey with AWS during the pandemic
Global video provider BlueJeans (a Verizon company) supports employees working from home, healthcare providers shifting to telehealth, and educators moving to distance learning. With so many people now working from home, BlueJeans saw explosive gro
Learn how healthcare organizations can harness the power of AI and machine learning to automate clinical workflows, digitize medical information, extract and summarize medical information, protect patient data, and more. Using AWS document understand
As clinical trials increasingly become decentralized and virtual, engaging effectively with patients can be challenging. In this session, hear how Evidation Health architects on AWS to create patient-centric experiences, ingests data from millions of devices in real time
BlueJeans’ explosive growth journey with AWS during the pandemic
Global video provider BlueJeans (a Verizon company) supports employees working from home, healthcare providers shifting to telehealth, and educators moving to distance learning. With so many people now working from home, BlueJeans saw explosive gro
Learn how healthcare organizations can harness the power of AI and machine learning to automate clinical workflows, digitize medical information, extract and summarize medical information, protect patient data, and more. Using AWS document understand
As clinical trials increasingly become decentralized and virtual, engaging effectively with patients can be challenging. In this session, hear how Evidation Health architects on AWS to create patient-centric experiences, ingests data from millions of devices in real time
Edge computing innovation with AWS Snowcone and AWS Snowball Edge
In this session, learn how the AWS Snow Family can help you run operations in harsh, non-data center environments and in locations where there is a lack of consistent network connectivity. The AWS Snow Family, comprised of AWS Snowcone and AWS Snowball Edge, offers a number of physical devices and capacity points with built-in computing capabilities. This sess… Learn More
Join AWS Healthcare and Life Science Leader Shez Partovi, MD, for a look into how cloud technology can reshape the future of healthcare. In this executive overview, Dr. Partovi shares a vision of a digitally enhanced, data-driven future. Learn how AWS is working
Edge computing innovation with AWS Snowcone and AWS Snowball Edge
In this session, learn how the AWS Snow Family can help you run operations in harsh, non-data center environments and in locations where there is a lack of consistent network connectivity. The AWS Snow Family, comprised of AWS Snowcone and AWS Snowball Edge, offers a number of physical devices and capacity points with built-in computing capabilities. This sess… Learn More
Join AWS Healthcare and Life Science Leader Shez Partovi, MD, for a look into how cloud technology can reshape the future of healthcare. In this executive overview, Dr. Partovi shares a vision of a digitally enhanced, data-driven future. Learn how AWS is worki
Edge computing innovation with AWS Snowcone and AWS Snowball Edge
In this session, learn how the AWS Snow Family can help you run operations in harsh, non-data center environments and in locations where there is a lack of consistent network connectivity. The AWS Snow Family, comprised of AWS Snowcone and AWS Snowball Edge, offers a number of physical devices and capacity points with built-in computing capabilities. This sess… Learn More
Join AWS Healthcare and Life Science Leader Shez Partovi, MD, for a look into how cloud technology can reshape the future of healthcare. In this executive overview, Dr. Partovi shares a vision of a digitally enhanced, data-driven future. Learn how AWS is worki
Intelligent document processing for the insurance industry
Organizations in the insurance industry, both for healthcare and financial services, extract sensitive information such as names, dates, claims, or medical procedures from scanned images and documents to perform their business operations. These organization
This session is open to anyone, but it is intended for current and potential AWS Partners. The COVID-19 pandemic has been a formative event affecting our world physically, emotionally, and economically. Despite the challenges created, AWS Partners have responded quickly and proven their resiliency by enabling digital transformation at unprecedented rates. In this sessi… Learn More
Add To Calendar
Global Partner Summit (GPS) Session
DEC 15, 2020 | 9:00 PM – 9:30 PM EST
Intelligent document processing for the insurance industry
Organizations in the insurance industry, both for healthcare and financial services, extract sensitive information such as names, dates, claims, or medical procedures from scanned images and documents to perform their business operations. These organization
This session is open to anyone, but it is intended for current and potential AWS Partners. The COVID-19 pandemic has been a formative event affecting our world physically, emotionally, and economically. Despite the challenges created, AWS Partners have responded quickly and proven their resiliency by enabling digital transformation at unprecedented rates. In this sessi… Learn More
Add To Calendar
Global Partner Summit (GPS) Session
Wednesday, December 16
DEC 16, 2020 | 5:00 AM – 5:30 AM EST
Intelligent document processing for the insurance industry
Organizations in the insurance industry, both for healthcare and financial services, extract sensitive information such as names, dates, claims, or medical procedures from scanned images and documents to perform their business operations. These organization
This session is open to anyone, but it is intended for current and potential AWS Partners. The COVID-19 pandemic has been a formative event affecting our world physically, emotionally, and economically. Despite the challenges created, AWS Partners have responded quickly and proven their resiliency by enabling digital transformation at unprecedented rates. In this sessi… Learn More
Add To Calendar
Global Partner Summit (GPS) Session
DEC 16, 2020 | 11:45 AM – 12:15 PM EST
Simplify healthcare compliance with third-party solutions
Sensitive health data must be protected to ensure patient privacy, and healthcare organizations need to ensure that IT infrastructure is compliant with changing policies and regulations. In this session, learn how healthcare providers can address compli
An introduction to healthcare interoperability and FHIR Works on AWS
Fast Healthcare Interoperability Resources (FHIR) is gaining popularity around the world as the standard to use for exchanging healthcare data, and it is being increasingly adopted in Europe and Australasia. In the US, it is actually mandated in the 21st Century
Achieving healthcare interoperability with FHIR Works on AWS
The Fast Healthcare Interoperability Resources (FHIR) standard has become increasingly necessary for enabling interoperability between healthcare applications and organizations. Join this session for a deep dive into how the FHIR Works on AWS open-source t
AWS at the edge: Using AWS IoT to optimize Amazon wind farms
AWS IoT and edge computing solutions move data processing and analysis closer to where data is generated to enable customers to innovate and achieve more sustainable operations. In this session, learn how Amazon renewable energy projects use AWS IoT to coll
Simplify healthcare compliance with third-party solutions
Sensitive health data must be protected to ensure patient privacy, and healthcare organizations need to ensure that IT infrastructure is compliant with changing policies and regulations. In this session, learn how healthcare providers can address compli
An introduction to healthcare interoperability and FHIR Works on AWS
Fast Healthcare Interoperability Resources (FHIR) is gaining popularity around the world as the standard to use for exchanging healthcare data, and it is being increasingly adopted in Europe and Australasia. In the US, it is actually mandated in the 21st Century
Achieving healthcare interoperability with FHIR Works on AWS
The Fast Healthcare Interoperability Resources (FHIR) standard has become increasingly necessary for enabling interoperability between healthcare applications and organizations. Join this session for a deep dive into how the FHIR Works on AWS open-source t
AWS at the edge: Using AWS IoT to optimize Amazon wind farms
AWS IoT and edge computing solutions move data processing and analysis closer to where data is generated to enable customers to innovate and achieve more sustainable operations. In this session, learn how Amazon renewable energy projects use AWS IoT to coll
Simplify healthcare compliance with third-party solutions
Sensitive health data must be protected to ensure patient privacy, and healthcare organizations need to ensure that IT infrastructure is compliant with changing policies and regulations. In this session, learn how healthcare providers can address compliance
An introduction to healthcare interoperability and FHIR Works on AWS
Fast Healthcare Interoperability Resources (FHIR) is gaining popularity around the world as the standard to use for exchanging healthcare data, and it is being increasingly adopted in Europe and Australasia. In the US, it is actually mandated in the 21st Century
Achieving healthcare interoperability with FHIR Works on AWS
The Fast Healthcare Interoperability Resources (FHIR) standard has become increasingly necessary for enabling interoperability between healthcare applications and organizations. Join this session for a deep dive into how the FHIR Works on AWS open-source t
AWS at the edge: Using AWS IoT to optimize Amazon wind farms
AWS IoT and edge computing solutions move data processing and analysis closer to where data is generated to enable customers to innovate and achieve more sustainable operations. In this session, learn how Amazon renewable energy projects use AWS IoT to coll
Share information by removing language barriers with Amazon Translate
In this session, see how to use Amazon Translate and Amazon Transcribe to create subtitles for educational videos and translate them to the language of the consumer’s choice. The session includes a demonstration of how this process reduced the time it took for information about COVID-19 to be translated to many languages, thus spreading accurate information quickly.
Share information by removing language barriers with Amazon Translate
In this session, see how to use Amazon Translate and Amazon Transcribe to create subtitles for educational videos and translate them to the language of the consumer’s choice. The session includes a demonstration of how this process reduced the time it took for information about COVID-19 to be translated to many languages, thus spreading accurate information quickly.
Share information by removing language barriers with Amazon Translate
In this session, see how to use Amazon Translate and Amazon Transcribe to create subtitles for educational videos and translate them to the language of the consumer’s choice. The session includes a demonstration of how this process reduced the time it took for information about COVID-19 to be translated to many languages, thus spreading accurate information quickly.
Share information by removing language barriers with Amazon Translate
In this session, see how to use Amazon Translate and Amazon Transcribe to create subtitles for educational videos and translate them to the language of the consumer’s choice. The session includes a demonstration of how this process reduced the time it took for information about COVID-19 to be translated to many languages, thus spreading accurate information quickly.
Bookmark this blog and check back for direct links to each session and add to your re:Invent schedule as soon as the session catalog is released:
LFS201: Life Sciences Industry: Executive Outlook
Learn how AWS technology is helping organizations improve their data liquidity, achieve operational excellence, and enhance customer engagement.
LFS202: Improving data liquidity in Roche’s personalized healthcare platform
Learn how Roche’s personalized healthcare platform is accelerating drug discovery and transforming the patient journey with digital technology.
LFS302: AstraZeneca genomics on AWS: from petabytes to new medicines
Learn how AstraZeneca built an industry leading genomics pipeline on AWS to analyze 2 million genomes in support of precision medicine.
LFS303: Building patient-centric virtualized trials
Learn how Evidation Health architects on AWS to create patient-centric experiences in decentralized and virtual clinical trials.
LFS304: Streamlining manufacturing and supply chain at Novartis
Learn how Novartis is creating real-time analytics and transparency in the pharma manufacturing process and supply chain to bring innovative medicines to market.
LFS305: Accelerating regulatory assessments in life sciences manufacturing
Learn how Merck leveraged Amazon Machine Learning to build an evaluation and recommendation engine for streamlining pharma manufacturing change requests.
Other related sessions of interest:
ENT203: How BMS automates GxP compliance for SAP systems on AWS
HLC203: Securing Personal Health Information and High Risk Data Sets
WPS202: Transform research environments with Service Workbench on AWS
AIM310: Intelligent document processing for healthcare organizations
Healthcare Attendee Guide
AWS re:Invent routinely fills several Las Vegas venues with standing-room only crowds, but we are bringing it to you with an all-virtual and free event this year. This year’s conference is gearing up to be our biggest yet and we have an exciting program planned for the Healthcare industry with five keynotes, 18 leadership sessions, and over 500 breakout sessions beginning November 30. See how AWS experts and talented members of the Healthcare industry are using cloud technology to transform their businesses and innovate on the behalf of their customers.For Healthcare attendees looking to get the most out of their experience, follow these steps:
Take a look at all of the Healthcare sessions available, as well as lots of other information and additional activities, in our curated Healthcare Attendee Guide coming soon!
Check back on this post regularly, as we’ll continually update it to reflect the newest information.
Healthcare at re:Invent 2020
AWS is the trusted technology partner to the global healthcare industry. For over 12 years, AWS has established itself as the most mature, comprehensive, and broadly adopted cloud platform and is trusted by thousands of healthcare customers around the world—including the fastest-growing startups, the largest enterprises, and leading government agencies. The secure and compliant AWS technology enables the highly regulated healthcare industry to improve outcomes and lower costs by providing the tools to unlock the potential of healthcare data, predict healthcare events, and build closer relationships with patients and consumers. The healthcare track at re:Invent 2020 will feature customer-led sessions focused on these each of these critical components, accelerating the transformation of healthcare.
Healthcare sessions
Learn more and bookmark each Healthcare session:
HCL201: Healthcare Executive Outlook: Accelerating Transformation
Learn how AWS is working with industry leaders to increase their pace of innovation, unlock the potential of their healthcare data, help predict patient health events, and personalize the healthcare journey for their patients, consumers, and members.
HLC202: Making Healthcare More Personal with MetroPlus Health
Learn how MetroPlus Health leveraged AWS technology to quickly build and deploy an application that personally and proactively reached out to its members during a time of critical need.
HLC203: Securing Personal Health Information and High Risk Data Sets
Learn how Arcadia developed a HITRUST CSF Certified platform by leveraging AWS technology to enable the secure management of data from over 100 million patients.
HLC204: Accelerating the Transition to Virtual Care with AWS
Learn how MedStar Health developed and deployed two call centers in less than week that are supporting more than 3,500 outpatient telehealth sessions a day.
WPS202: Transform research environments with Service Workbench on AWS
Learn how Harvard Medical School is using AWS to procure and deploy domain-specific data, tools, and secure IT environments to accelerate research.
WPS209: Reinventing medical imaging with machine learning on AWS
Learn how Radboud University Medical Center uses AWS to power its machine learning imaging platform with 45,000+ registered researchers and clinicians from all over the world.
WPS211: An introduction to healthcare interoperability and FHIR Works on AWS
Learn about AWS FHIR Works, an open-source project, designed to accelerate the industries use of the interoperability standard, Fast Healthcare Interoperability Resources (FHIR).
WPS304: Achieving healthcare interoperability with FHIR Works on AWS
Learn how Black Pear Software leveraged AWS to build an integration toolkit to help their customers share healthcare data more effectively.
Extras you won’t want to miss out on!
LFS201: Life Sciences Industry: Executive outlook
LFS202: AstraZeneca genomics on AWS: From petabytes to new medicines
LFS303: Building patient-centric virtualized trials
AIM303: Using AI to automate clinical workflows
AIM310: Intelligent document processing for the insurance industry
INO204: Solving societal challenges with digital innovation on AWS
ZWL208: Using cloud-based genomic research to reduce health care disparities
Ahmet K. Karagozoglu, Ph.D. | C.V. Starr Distinguished Professor of Finance | Visiting Scholar, Volatility and Risk Institute | Hofstra University | New York University Stern School of Business
Type: Track Keynote, Level: All Levels, Focus Area: NLP
Intelligibility Throughout the Machine Learning Life Cycle
Sveta Kostinsky | Director of Sales Engineering | Samasource
Marcelo Benedetti | Senior Account Executive | Samasource
Type: Talk, Level: Intermediate, Focus Area: Machine Learning, Deep Learning
quality rubric
Internal QA Sampling
Client QA Sampling
Auto QA
Transfer Learning in NLP
Thu, October 29, 3:40 PM (PDT)
00:
03:
30
Joan Xiao, PhD | Principal Data Scientist | Linc Global
Type: Talk, Level: Intermediate, Focus Area: NLP, Deep Learning
Transfer learning enables leveraging knowledge acquired from related data to improve performance on a target task. The advancement of deep learning and large amount of labelled data such as ImageNet has made high performing pre-trained computer vision models possible. Transfer learning, in particular, fine-tuning a pre-trained model on a target task, has been a far more common practice than training from scratch in computer vision.
In NLP, starting from 2018, thanks to the various large language models (ULMFiT, OpenAI GPT, BERT family, etc) pre-trained on large corpus, transfer learning has become a new paradigm and new state of the art results on many NLP tasks have been achieved.
In this session we’ll learn the different types of transfer learning, the architecture of these pre-trained language models, and how different transfer learning techniques can be used to solve various NLP tasks. In addition, we’ll also show a variety of problems that can be solved using these language models and transfer learning.
Transfer learning: Computer Vision – ImageNet Classification
OpenAI built a text generator – too dangerous to release
OpenAI GPT-3 – Trained on 300B tokens – THREE models:
Zero-shot – English to French – no training
one-shots
Few-shot – the GOAL – GPT-3
GRT-3 is large scale NLP
Examples – Feature extraction
English to SQL
English to CSS
English to LaTex
Semantic textual similarity
NL inference
ULMFiT – Fine tuning – the larger the # of Training examples – the better the performance
LM pre-training – start from scratch: BART, Big Bird, ELECTRA, Longformer
LM fine-tuning
Classifier fine-tuning
Data augmentation
Contextual Augmentation
Original sentence
masked
augmented
Test generation
boolean questions
from structured data, i.e., RDF – Resource Description Framework
OCTOBER 30TH
Generalized Deep Reinforcement Learning for Solving Combinatorial Optimization Problems
Fri, October 30, 9:00 AM (PDT)
Azalia Mirhoseini, PhD | Senior Research Scientist | Google Brain
Type: Keynote
Abstract:
Many problems in systems and chip design are in the form of combinatorial optimization on graph structured data. In this talk, I will motivate taking a learning based approach to combinatorial optimization problems with a focus on deep reinforcement learning (RL) agents that generalize. I will discuss our work on a new domain-transferable reinforcement learning methodology for optimizing chip placement, a long pole in hardware design. Our approach is capable of learning from past experience and improving over time, resulting in more optimized placements on unseen chip blocks as the RL agent is exposed to a larger volume of data. Our objective is to minimize PPA (power, performance, and area), and we show that, in under 6 hours, our method can generate placements that are superhuman or comparable on modern accelerator chips, whereas existing baselines require human experts in the loop and can take several weeks.
Bio:
Azalia Mirhoseini is a Senior Research Scientist at Google Brain. She is the co-founder/tech-lead of the Machine Learning for Systems Team in Google Brain where they focus on deep reinforcement learning based approaches to solve problems in computer systems and metal earning. She has a Ph.D. in Electrical and Computer Engineering from Rice University. She has received a number of awards, including the MIT Technology Review 35 under 35 award, the Best Ph.D. Thesis Award at Rice and a Gold Medal in the National Math Olympiad in Iran. Her work has been covered in various media outlets including MIT Technology Review, IEEE Spectrum, and Wired.
Placement Optimmization using AGENTS to place the nodes
Train Policy to be using for placement of ALL chips
Compiling a Dataset of Chip Placements
Policy/Value Model Architecture to save wire length used
RISC-V: Placement Visualization: Training from Scratch (Human) 6-8 weeks vs Pre-Trained 24 hours
Keynote Session – Zoubin Ghahramani
Fri, October 30, 9:30 AM (PDT)
Zoubin Ghahramani, PhD | Distinguished Scientist and Sr Research Director | Professor of Information Engineering | ex-Chief Scientist and VP of AI | Google | University of Cambridge | Uber
Automating Augmenting Data with knowledge: feature-concept mapping
Modeling
Time Series Forecasting
AI to decision Optimization
Demand forecasting from Standard AutoAI by ADDING Historical Decisions and Historical Business Impact__>> reinforced learning – Automatically created model from past and Auto AI
Validation
Meta-learning for performance prediction
Train the META data
Score production data with AI
Deployment
staged deployment with contextual bandits
Monitoring
Performance prediction meta model applied over windows of production traffic
INNOVATIONS;
End-to-end AI life cycle
expanding scope of automation; Domain knowledge and decision optimization
The State of Serverless and Applications to AI
Fri, October 30, 11:20 AM (PDT)
Joe Hellerstein, PhD | Chief Strategy Officer, Professor of Computer Science | Trifacta, Berkeley
The Cloud and practical AI have evolved hand-in-hand over the last decade. Looking forward to the next decade, both of these technologies are moving toward increased democratization, enabling the broad majority of developers to gain access to the technology.
Serverless computing is a relatively new abstraction for democratizing the task of programming the cloud at scale. In this talk I will discuss the limitations of first-generation serverless computing from the major cloud vendors, and ongoing research at Berkeley’s RISELab to push forward toward “”””stateful”””” serverless computing. In addition to system infrastructure, I will discuss and demonstrate applications including data science, model serving for machine learning, and cloud-bursted computing for robotics.
Bio:
Joseph M. Hellerstein is the Jim Gray Professor of Computer Science at the University of California, Berkeley, whose work focuses on data-centric systems and the way they drive computing. He is an ACM Fellow, an Alfred P. Sloan Research Fellow and the recipient of three ACM-SIGMOD “Test of Time” awards for his research. Fortune Magazine has included him in their list of 50 smartest people in technology , and MIT’s Technology Review magazine included his work on their TR10 list of the 10 technologies “most likely to change our world”. Hellerstein is the co-founder and Chief Strategy Officer of Trifacta, a software vendor providing intelligent interactive solutions to the messy problem of wrangling data. He has served on the technical advisory boards of a number of computing and Internet companies including Dell EMC, SurveyMonkey, Captricity, and Datometry, and previously served as the Director of Intel Research, Berkeley.
Type: Talk, Level: Intermediate, Focus Area: AI for Good, Machine Learning
Tina Eliassi-Rad, PhD | Professor | Core Faculty | Northeastern University | Network Science Institute
Type: Talk, Level: All Levels, Focus Area: Machine Learning
In 1997, Tom Mitchell defined the well-posed learning problem as follows: “A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E.” In this talk, I will discuss current tasks, experiences, and performance measures as they pertain to fairness in machine learning. The most popular task thus far has been risk assessment. We know this task comes with impossibility results (e.g., see Kleinberg et al. 2016, Chouldechova 2016). I will highlight new findings in terms of these impossibility results. In addition, most human decision-makers seem to use risk estimates for efficiency purposes and not to make fairer decisions. I will present an alternative task definition whose goal is to provide more context to the human decision-maker. The problems surrounding experience have received the most attention. Joy Buolamwini (MIT Media Lab) refers to these as the “under-sampled majority” problem. The majority of the population is non-white, non-male; however, white males are overrepresented in the training data. Not being properly represented in the training data comes at a cost to the under-sampled majority when machine learning algorithms are used to aid human decision-makers. In terms of performance measures, a variety of definitions exist from group- to individual- to procedural-fairness. I will discuss our null model for fairness and demonstrate how to use deviations from this null model to measure favoritism and prejudice in the data.
Tasks:
Assessing risk
Ranking
Statistical parity: among classifier
PARITY vs imperfect classifier – can’t satisfy all the three conditions
Precision
Tru positive
False parity
All classifier do not consider context or allow for uncertainty
Learning to Place within existing cases
Incentives/values of Human decision maker which incorporate in the decision external factors
Game-theoretical framework
How human exemplars make decision
Are algorithms value free?
Computational Ethics
Logically consistent principle
Camouflage – machine did not learn on the task but on the cloudiness of the sky
Model Cards for Model Reporting
The “undersampled majority”
Experience: Demonstration: Should we learn from demonstrations or from simulations?
Complex networks: guilt by association vs privilege and prejudice, individual fairness
Datasheets for Datasets
Algorithms are like prescription drug: Adverse events
Human vs Machine judgement
Performance measure – FAIRNESS: Group, individual
Normativity throughout the entire well-posed learning problem
Incentive/values
Human or machines to make decisions?
Laws are needed if algorithms are used as expert witness
Machine Learning for Biology and Medicine
Fri, October 30, 2:00 PM
Sriram Sankararaman, PhD | Professor, Computer Science | University of California – Los Angeles
Type: Talk, Focus Area: Machine Learning
Abstract:
Biology and medicine are deluged with data so that techniques from machine learning and statistics will increasingly play a key role in extracting insights from the vast quantities of data being generated. I will provide an overview of the modeling and inferential challenges that arise in these domains.
In the first part of my talk, I will focus on machine learning problems arising in the field of genomics. The cost of genome sequencing has decreased by over 100,000 fold over the last decade. Availability of genetic variation data from millions of individuals has opened up the possibility of using genetic information to identifying the cause of diseases, developing effective drugs, predicting disease risk and personalizing treatment. While genome-wide association studies offer a powerful paradigm to discovering disease-causing genes, the hidden genetic structure of human populations can confound these studies. I will describe statistical models that can infer this hidden structure and show how these inferences lead to novel insights into the genetic basis of diseases.
In the second part of my talk, I will discuss how the availability of large-scale electronic medical records is opening up the possibility of using machine learning in clinical settings. These electronic medical records are designed to capture a wide range of data associated with a patient including demographic information, laboratory tests, images, medications and clinical notes. Using electronic records from around 60,000 surgeries over five years in the UCLA hospital, I will describe efforts to use machine learning algorithms to predict mortality after surgery. Our results reveal that these algorithms can accurately predict mortality from information available prior to surgery indicating that automated predictive systems have great potential to augment clinical care.
Bio:
Sriram Sankararaman is an assistant professor in the Departments of Computer Science, Human Genetics, and Computational Medicine at UCLA where he leads the machine learning and genomic lab. His research interests lie at the interface of computer science, statistics and biology and is interested in developing statistical machine learning algorithms to make sense of large-scale biomedical data and in using these tools to understand the interplay between evolution, our genomes and traits. He received a B.Tech. in Computer Science from the Indian Institute of Technology, Madras, a Ph.D. in Computer Science from UC Berkeley and was a post-doctoral fellow in Harvard Medical School before joining UCLA. He is a recipient of the Alfred P. Sloan Foundation fellowship (2017), Okawa Foundation grant (2017), the UCLA Hellman fellowship (2017), the NIH Pathway to Independence Award (2014), a Simons Research fellowship (2014), and a Harvard Science of the Human Past fellowship (2012) as well as the Northrop-Grumman Excellence in Teaching Award at UCLA (2019).
ML & BioMedicine
BioMedical data: high D, heterogeneous, noisy data
Clinical Data & DL
Predict death after surgery – 1000 dealth complication, sepsis acout kidney injury
Mortality during and after surgery
collaboration: Anesthesiology, PeriOps, UCLA Health
Data warehouse – EMR 4/2013 – 12/2018
60,000 patients in data: Age, height, weight, gender,ASA Status- input from physician
Pre-operative mortality risk prediction – False positive, missing data: Lab data was collected, what were the values
With the computational advances over the past few decades, Bayesian analysis approaches are starting to be fully appreciated. Forecasting and time series also have Bayesian approaches and techniques, but most people are unfamiliar with them due to the immense popularity of Exponential Smoothing and autoregressive integrated moving average (ARIMA) classes of models. However, Bayesian modeling and time series analysis have a lot in common! Both are based on using historical information to help inform future modeling and decisions. Using past information is key to any time series analysis because the data typically evolves over time in a correlated way. Bayesian techniques rely on new data updating their models from previous instances for better estimates of posterior distributions. This talk will briefly introduce the differences between classical frequentist approaches of statistics to their Bayesian counterparts as well as the difference between time series data made for forecasting compared to traditional cross-sectional data. From there, it will compare the classical Exponential Smoothing and ARIMA class models of time series to Bayesian models with autoregressive components. Comparing the results of these models across the same data set allows the audience to see the potential benefits and disadvantages of using each of the techniques. This talk aims to allow people to update their own skill set in forecasting with these potentially Bayesian techniques. At the end, the talk explores the technique of model ensembling in a time series context. From these ensembles, the benefits of all types of models are potentially blended together. These models and their respective outputs will be displayed in R
Single Exponential Smoothing
ARIMA – long-memory models – Autoregressive AR
Moving Average (MA) model – short memory
Intergrated AR+MA = ARIMA
Learning Intended Reward Functions: Extracting all the Right Information from All the Right Places
Fri, October 30, 3:45 PM
(PDT)
00:04:42
Anca Dragan, PhD | Assistant Professor, EECS | Head | UC Berkeley | InterACT lab
Learning Intended Reward Functions: Extracting all the Right Information from All the Right Places
Abstract:
Content: AI work tends to focus on how to optimize a specified reward function, but rewards that lead to the desired behavior consistently are not so easy to specify. Rather than optimizing specified reward, which is already hard, robots have the much harder job of optimizing intended reward. While the specified reward does not have as much information as we make our robots pretend, the good news is that humans constantly leak information about what the robot should optimize. In this talk, we will explore how to read the right amount of information from different types of human behavior — and even the lack thereof.
Learning outcomes: After participating, you should be able to articulate the common pitfalls we face in defining an AI reward, loos, or objective function. You should also develop a basic understanding of the main algorithmic tools we have for avoiding these pitfalls.
Target audience: Participants with some AI experience, be in supervised or reinforcement learning.
Bio:
Anca Dragan is an Assistant Professor in EECS at UC Berkeley, where she runs the InterACT lab. Her goal is to enable robots to work with, around, and in support of people. She works on algorithms that enable robots to a) coordinate with people in shared spaces, and b) learn what people want them to do. Anca did her PhD in the Robotics Institute at Carnegie Mellon University on legible motion planning. At Berkeley, she helped found the Berkeley AI Research Lab, is a co-PI for the Center for Human-Compatible AI, and has been honored by the Presidential Early Career Award for Scientists and Engineers (PECASE), the Sloan fellowship, the NSF CAREER award, the Okawa award, MIT’s TR35, and an IJCAI Early Career Spotlight.
Sequential decision making
defining what robots goal is
Autonomous car
AI = optimize intended rewards vs specified reward
parametrization of the reward function
Agent over-learn from specified rewards but under-learn from other sources
observing feedback and express the human feedback in observation (human) model
How can we model reward design/specification as a noisy and suboptiman process
Human feedback as a reward-rational implicit choice
The state of the environment as a reward-rational implicit choice
task specification –>> reward
KEYNOTE SPEAKERS
ODSC West Keynotes
Suchi Saria, PhD
Director Of The Machine Learning And Healthcare Lab, John C. Malone Endowed Chair, Founder Of Bayesian Health, MIT Technology Review’s 35 Innovators Under 35, And A World Economic Forum Young Global Leader
Johns Hopkins University
Jeannette M. Wing, PhD
Avanessians Director Of The Data Science Institute, Professor Of Computer Science Columbia University, Former Corporate Vice President Microsoft, Former Assistant Director, National Science Foundation
Columbia University
Ion Stoica, PhD
Professor Of Computer Science, Head Of RISELab. Co-Founder Of Anyscale, Databricks, And Conviva Networks, ACM Fellow, SIGOPS Hall Of Fame Award (2015), SIGCOMM Test Of Time Award (2011)
UC Berkeley
Raluca Ada Popa, PhD
Cybersecurity & Applied Cryptography Professor, MIT Technology Review’s 35 Under 35, Recipient Of Intel Early Career Faculty Honor Award, George M. Sprowls Award For Best MIT CS Doctoral Thesis, Co-Founder Of PreVeil
UC Berkeley
Zoubin Ghahramani, PhD
Chief Scientist, Founding Director Of The AlanTuring Institute, Prof. Of Information Engineering & Deputy Director Of The Leverhulme Centre For The Future Of Intelligence, Fellow Of St John’s College Cambridge And Of The Royal Society
Uber | The University of Cambridge
Azalia Mirhoseini, PhD
Senior Research Scientist At Google Brain. Advisor At Cmorq. Co-Founder Machine Learning For Systems Moonshot At Brain Focusing On Deep RL. MIT Technology Review 35 Under 35 Award
Google Brain
Marinela Profi
Global Strategist For AI, Global Ambassador For The Women Tech Network, Author Of “Mastering Model Lifecycle Orchestration: An Interactive Guide”
SAS
John Montgomery
Corporate Vice President, Visual Studio, Microsoft Azure AI Lead, Former Chief Information Office At Imagine Publishing, Author At Visual Studio
Microsoft
Ben Taylor,PhD
Chief AI Evangelist, Deep Learning & HPC Expert, Co-Founder & Chief Scientist At Zeff.Ai, Former Chief Scientist At HireVue, ProductCraft Contributor
Julian Hong, MD, MS
Charlotta Lindvall, MD, PhD
Tina Hernandez-Boussard, PhD
Imon Banerjee, PhD
2012pharmaceutical
Amandeep Kaur
Aashir Awan, Phd
Adina Hazan
Alan F. Kaul, PharmD., MS, MBA, FCCP
alexcrystal6
anamikasarkar
apreconasia
aviralvatsa
David Orchard-Webb, PhD
danutdaagmailcom
Demet Sag, Ph.D., CRA, GCP
Daniel Menzin
Dror Nir
dsmolyar
evelinacohn
Gail S Thornton
Irina Robu
jayzmit48
jdpmd
jshertok
kellyperlman
Ed Kislauskis
larryhbern
Madison Davis
marzankhan
megbaker58
ofermar2020
Dr. Pati
pkandala
Rosalind Codrington, PhD
ritusaxena
Rick Mandahl
sjwilliamspa
stuartlpbi
Dr. Sudipta Saha
tildabarliya
zraviv06
zs22