LIVE – Real Time Coverage of Kailath Lecture and Colloquium 2015 September 17 ‐ 19, Stanford CA – James H. Clark Center, 318 Campus Drive West, Stanford University, CA
Reporter: Aviva Lev-Ari, PhD, RN
Thursday September 17
9:00-9:30 AM:
Introduction: Profs Roychowdhury and Kailath
9:30-‐11am: Prof Roychowdhury, Chair
- Freddy Bruckstein (Technion): From Ants to A(ge)nts
-
Ali Sayed (UCLA): How well can we learn over networks?
-
George Verghese (MIT): Physiological Modeling and Estimation for Bedside Informatics
11:30 – 12:30 PM: Prof. Varvarigou, Chair
- H. V. Jagadish (Michigan: The Ethics of Big Data
- S.Y.Kung (Princeton) Visualization of Big Data
1:30 – 3:30PM Prof. Orlitsky, Chair
- Vwani Roychowdhury (UCLA): What can Aliens infer about Humans from the Internet?
- Theodora Varvarigou (NTU Athens): Edge-Sign Prediction in Social Networks using Frequent subgraph Discovery
Homophily (similarity) ->> Social Influence –>> Trust
Canonical Labeling
Performance improvement – achieved
- Lennart Ljung (Linkoping University, Sweden): Will Machine Learning Change the Paradigm of System Identification?
TK – intuitive attitude to Mathematics, was part of Birthday 50th, 60th, 70th and 80th
System identification: Model Structure: Maximum likelihood THEN Select the Order best: Cross Validation
Function Learning: Gaussian Process Regression & Hilbert Space reproduced by the Kernel of the Space
Estimation: Bias-Variance-Tradeoff: Regularazation matrix
Empirical Bayes: Stable splines
- Helmut Bolcskei (ETH, Zurich) Super-resolved system identification
Connectivity – General bound linear operators: Spreading function, kernel
Channel sounding
Control engineering
Radar imaging
Partial systems are sparce – spectrum-blind sampling
Identification Algorithms:
4:00—5:30pm: Kailath Lecture, Prof Goldsmith, Chair
Welcome: Prof. El Gamal (EE Chair) and Prof. Kailath;
TK: recipient of National Medal of Science Award in 2014
Mentor, colleague, warm and supportive friend,
Speaker to be introduced by Prof. Yablonovitch (UCB)
- sub wavelet Imaging Joint collaboration Prof. Yablonovitch (UCB) and Prof. Stanley Osher (UCLA)
Prof. Stanley Osher (UCLA): What Mathematical Algorithms Can Do for the Real (and Even Fake) World
Dicontinuities developing spontaneously
Solutions: Essential Non-oscillatory Interpolations (ENO) SMOOTHING – useful in Image Processing
EDGING ARE FOUNDAMENTAL IN IMAGING — Shockwaves and Imaging
linear and quadratic
Problems arise e.g. in Control Theory and Dynamic Programming
Leave a Reply