Science Videos and Lectures

How a tiny change in a molecule’s geometry completely changes its effects on the human body.

Time series appear in a variety of key real-world applications such as signal processing, including audio and video processing; the analysis of natural phenomena such as local weather, global temperature, and earthquakes; the study of economic variables such as stock values, sales amounts, energy demand; and many other areas. But, while time series forecasting is critical for many applications, it has received little attention in the ML community in recent years, probably due to a lack of familiarity with time series and the fact that standard independent and identically distributed learning concepts and tools are not readily applicable in that scenario.

This tutorial precisely addresses these and many other related questions. It provides theoretical and algorithmic tools for research related to time series and for designing new solutions. We first present a concise introduction to time series, including basic concepts, common challenges and standard models. Next, we discuss important statistical learning tools and results developed in recent years and show how they are useful for deriving guarantees and designing algorithms both in stationary and non-stationary scenarios. Finally, we show how the online learning framework can be leveraged to derive algorithms that tackle important and notoriously difficult problems including model selection and ensemble methods.

Prof. Arthur Jaffe (Landon T. Clay Professor of Mathematics and Theoretical Science at Harvard University) recently gave the Statutory Public Lecture of the School of Theoretical Physics, DIAS on "Beauty and Truth in Mathematics and Physics." The lecture was held on Wednesday 18th May in the Schrodinger Lecture Theatre at Trinity College Dublin. Professor Jaffe is also Chair of the Governing Board of the School of Theoretical Physics, DIAS.

Dozens of playlists consisting of multiple ~5-10 minute videos on many different subjects within machine learning, from basic clustering, regression, and probability algorithms to neural networks, information retrieval, search engines, and text analysis by Victor Lavrenko of the University of Edinburgh.

My name’s Allison Parrish, and I have a little talk prepared here. The title here is on the screen, “Programming is Forgetting: Toward a new hacker ethic.” I love that word “toward,” because it instantly makes any talk feel super serious. So I’m just going to begin.

A lecture on quantum information and spacetime by John Preskill, the Richard P. Feynman Professor of Theoretical Physics at CalTech.

The Thirtieth Annual Conference on Neural Information Processing Systems (NIPS) is a multi-track machine learning and computational neuroscience conference that includes invited talks, demonstrations, symposia and oral and poster presentations of refereed papers. Following the conference, there are workshops which provide a less formal setting.

A workshop linking the past, present and future research on neural networks held in New York, NY in 2016. The workshop homepage contains links to videos, papers, slides, and related code from the lectures.

In this Web of Stories video, scientist Marvin Minsky shows off his neural network machine.

DARPAtv is the Defense of Advanced Research Projects Agency (DARPA) YouTube Channel featuring videos on the latest scientific results from DARPA-funded researchers, lectures from DARPA administrators, the "Voices from DARPA" Podcast, and more.

Videos about scientific research and development at the Grenoble Innovation for Advanced New Technologies (GIANT) in Grenoble, France.

A tutorial lecture by Sherry Moore of Google Brain on TensorFlow from the Bay Area Deep Learning School (http://www.bayareadlschool.org) held on September 24 and 25, 2016.

Lex Friedman clipped out individual talks from the full live streams of the Bay Area Deep Learning School (http://www.bayareadlschool.org) held on September 24/25, 2016 and provided links to each lecture in the video summary. This one by Hugo Larochelle is the first lecture in the series. Other speakers include Andrej Karpathy, Richard Socher, Andrew Ng, Quoc Le, Yoshua Bengio, and more.

Videos in this series explain the intuition behind exciting cutting edge research results in an understandable way, a couple minutes at a time. We are mostly exploring the area of computer graphics and machine learning with a hint of philosophy, physics and medicine. You'll see how exciting these works are and how they shape our everyday lives.

Find presentations by resident and visiting scientists and researchers in theoretical physics given at PI.

How a certain perspective on what the Riemann zeta function looks like can motivate what it might mean beyond its domain of convergence.

An unsolved conjecture, the inscribed square problem, and a clever topological solution to a weaker version of the question, the inscribed rectangle problem, that shows how the torus and mobius strip naturally arise in mathematical ponderings. Proof due to H. Vaughan (1977).

The JR is a seagoing research vessel that drills core samples and collects measurements from under the ocean floor, giving scientists a glimpse into Earth’s development. Watch these videos to see how deep sea drilling and coring happens at sea.

48 video lectures on Python for data analysis and visualization from the PyData SF 2016 conference held on August 24, 2016.

Presented by Iain Murray, December 15, 2015. “Monte Carlo” methods use random sampling to understand a system, estimate averages, or compute integrals. Monte Carlo methods were amongst the earliest applications run on electronic computers in the 1940s, and continue to see widespread use and research as our models and computational power grow. In the NIPS community, random sampling is widely used within optimization methods, and as a way to perform inference in probabilistic models. Here “inference” simply means obtaining multiple plausible settings of model parameters that could have led to the observed data. Obtaining a range of explanations tells us both what we can and cannot know from our data, and prevents us from making overconfident (wrong) predictions.