About this channel
Microsoft Research offers a research-focused view of computing, with talks and workshops spanning generative models, biological applications, data structures and communication security. The episode…
Microsoft Research offers a research-focused view of computing, with talks and workshops spanning generative models, biological applications, data structures and communication security. The episode titles retain the vocabulary of the underlying work, including Gibbs sampling, convex potentials and probabilistic inference. That level of specificity makes the selection especially relevant to researchers, advanced students and technical viewers looking for the questions being investigated within a field.
Generative modelling and sampling form a substantial cluster. Workshop recordings sit alongside talks on the existence and limitations of one-shot generative flows and ways to control diffusion models through probabilistic inference. A discussion of generative Gibbs sampling introduces the role of explicit physical context. These subjects allow viewers to follow distinct research approaches within a shared area, with the mathematical formulation remaining visible in the way the talks are presented.
Applied work widens the picture. An introduction to Quine explores its role in AI research on the complexity of biology, while another episode explores AI agents for therapeutic reasoning across biological contexts. Elsewhere, a talk asks how AI changes power system planning. The selection therefore connects methodological research with questions arising in particular scientific and infrastructure settings.
Systems and security topics provide further routes in, including index trees designed to avoid thread switches and stalls, and an analysis of encryption in group messaging. Viewers can choose a specialised talk or spend longer with the workshop material. The channel provides a way to encounter research problems in their own technical terms and see how different areas of computing approach questions of modelling, scale and application.
Latest episode
What AI gets wrong and what failure teaches us
Jennifer Neville did not want to go into computer science—but that’s exactly where she landed. Neville discusses the starts and stops that led to her professional sweet spot and her work identifying…
Jennifer Neville did not want to go into computer science—but that’s exactly where she landed. Neville discusses the starts and stops that led to her professional sweet spot and her work identifying “surprising failures” making it hard for AI to handle complexity.
Show notes: https://www.microsoft.com/en-us/research/podcast/what-ai-gets-wrong-and-what-failure-teaches-us/
YouTube Shorts
Academic Research Summit 2018
Academic Research Summit 2019
Academic Research Summit 2020
AI and Social Good
AI and Societal Good Workshop 2018
AI Distinguished Lecture Series
AI Institute “Geometry of Deep Learning” 2019
Applied Multi-Party Computation
Artificial Intelligence (AI) Research
ASG Big Data Symposium 2015
Audio & Acoustics
Augmented Reality & Virtual Reality (AR & VR)
More episodes from this channel
Generative Gibbs sampling: Composing generative models with explicit physical context
1.10.2026
Generative models are powerful when training data for the target distribution are available, but in scientific computing the target is often a joint system for which no such data exist. We propose Generative Gibbs...
One-shot generative flows: Existence and obstructions
1.10.2026
In this talk, we study dynamic measure transport for generative modeling, focusing on transport maps that connect a source measure $P_0$ to a target measure $P_1$ by integrating a (conditional) velocity field of the...
The Future of the Grid: How AI Transforms Power System Planning
30.9.2026
Power grids are becoming increasingly complex as renewable energy, electrification, extreme weather, and growing demand reshape how energy systems operate. In this video, Microsoft Research introduces GridSFM, an open foundation model for power systems...
Latch-Free Index Trees That Scale: Avoiding Thread Switches & Stalls
28.9.2026
Speaker: David B. Lomet, Microsoft Research (retired) Host: Phil Bernstein, Microsoft Research Redmond Latch contention is a frequent impediment to high and scalable performance when accessing indexed data, even when I/O induced thread stalls...
Sample and Map from a Single Convex Potential: Generation using Conjugate Moment Measures
16.9.2026
The standard approach to generative modeling separates model fitting into two independent steps: first choosing a noise distribution to sample from, and then learning a transformation that maps these samples to the data distribution....
Probabilistic Inference for Controlling Diffusion Models
10.9.2026
Diffusion models are probabilistic models, and controlling a diffusion model at inference time is a probabilistic inference problem. In this talk, we will take a tour through inference-time control for diffusion models and organize...
MSRNE Generative Modeling & Sampling Workshop | Day 2
9.9.2026
The goal of the Generative Modeling & Sampling Workshop was to foster dialogue across communities working at the intersection of generative modeling, sampling, stochastic control, scientific machine learning, and related areas. By bringing together...
MSRNE Generative Modeling & Sampling Workshop | Day 1
9.9.2026
The goal of the Generative Modeling & Sampling Workshop was to foster dialogue across communities working at the intersection of generative modeling, sampling, stochastic control, scientific machine learning, and related areas. By bringing together...
AI agents for therapeutic reasoning across biological contexts
2.9.2026
Therapeutic hypotheses can transfer across diseases but their relevance depends on biological context. The same target, perturbation, or treatment can produce different effects across cell types, disease states, genetic backgrounds, and patients. Therapeutic reasoning...
Analyzing Group Chat Encryption in Messaging Applications
27.8.2026
Secure group messaging applications have been widely deployed to protect the conversations of billions of users. Several academic works have analyzed the cryptographic security of these applications, but most of them focus on analyzing...
Invited Research Talk: Measuring Generalization in EEG Foundation Models
27.8.2026
Over the past few years, there has been an increasing interest towards building EEG foundation models that provide generalized representations for EEG signals, which are able to provide noticeable performance improvements over supervised baseline...
TruthTable: A Verifiable Query Engine
27.8.2026
TruthTable is a verifiable database engine that allows a prover to produce a succinct proof that convinces a verifier of the correct execution of the verifier's SQL query over the prover's committed database. TruthTable...
AI, Radio Astronomy, and the Search for Life Beyond Earth
27.8.2026
Radio astronomy combines astronomy, physics, signal processing, engineering, and computing to study the Universe through radio-frequency observations. As modern radio telescopes generate vast volumes of data, AI is becoming an essential tool for advancing...
PhylaFlow: Hybrid flow matching in phylogenetic tree space
19.8.2026
Phylogenetic inference is a challenging generative modeling problem because phylogenetic trees combine continuous branch lengths with discrete topological structure. Bayesian methods provide a principled framework for representing uncertainty, but exploring the resulting posterior distribution...
Binomial Flows: Denoising and flow matching for discrete ordinal data
19.8.2026
Flow-based generative modeling in continuous spaces exploit Tweedie's formula to express the denoiser (learned in training) as a score function (used in sampling). In contrast, this relation has been largely missing in the discrete...
Expanding Flows for Fast and Flexible Generation Beyond the Fixed Canvas
22.7.2026
Flow-based generative models have enabled remarkable progress in fast and controllable generation across continuous and discrete state spaces, yet existing parameterizations are constrained to fixed dimensions or fixed sequence lengths. In this talk, I...
Microsoft AI for Good Lab - Introduction to HASTE
20.7.2026
Presenters: CVP Juan Lavista Ferres, Caleb Robinson, Cameron Birge, Kevin White The Microsoft AI for Good Lab introduces HASTE, its open-source rapid building-damage assessment tool that turns post-disaster satellite, aerial, or drone imagery into...
Learning Genetic Perturbation Effects at Single-Cell Resolution for Virtual Cells
16.7.2026
Complex causal mechanisms among genes govern cellular functions in health and disease. Understanding these mechanisms can accelerate therapeutic discovery but remains challenging due to the large number of genes and their intricate dependencies. Recent...
Convergence Analysis for Fast High-Order ODE Solvers in Diffusion Probabilistic Models
9.7.2026
Score-based generative models have revolutionized high-dimensional sampling through forward diffusion and reverse processes. While stochastic DDPM samplers benefit from mature polynomial convergence theory, deterministic probability flow ODEs (underlying efficient DDIM-style samplers) offer superior speed...
Human-Centered AI: Design, Deployment & Healthcare
7.7.2026
Microsoft Research India Academic Research Summit 2026 Speakers: Manik Gupta, BITS Hyderabad; Anirudha Joshi, IIT Bombay; Aaditeshwar Seth, IIT Delhi; Ashish Makhani, Ashoka University; Mohit Jain, Microsoft; Tavpritesh Sethi, IIIT Delhi #MSRI #ars2026 See...
Continue reading
You may also like
Innohub TV
Watch this content also on Innohub TV
We picked clips from Innohub TV that continue the same topic.
