180 days 1:56:28 Ultra Ethernet for next-generation AI and HPC workloads
Host: Sujata Banerjee Speakers: Torsten Hoefler, Microsoft and ETH Zurich; Abdul Kabbani, Microsoft The Ultra Ethernet Consortium set out to redefine Ethernet-based interconnects for AI and
2 597
13.6
180 days 53:58 BRAIN SIGNALS TO ACTION: Monitoring and Explaining User Cognitive Load with Foundation Models
Host: Dimitra Emmanouilidou, Microsoft Research Speaker: Deeksha Moodasarige Shama Passive monitoring of cognitive load can enable personalized user experiences and even accelerate human learning
728
9.7
180 days 34:11 IronDict: Transparent Dictionaries from Polynomial Commitments
Host: Melissa Chase, Microsoft Research Speaker: Hossein Hafezi, NYU We present IronDict, a transparent dictionary construction based on polynomial commitment schemes.
185
180 days 54:46 Lattice-Based Accumulator and Application to Anonymous Credential Revocation
Host: Betül Durak, Microsoft Research Speaker: Victor Youdom Kemmoe An accumulator is a cryptographic system for compactly representing a set of elements such that every element in the set has a
235
180 days 54:08 FOA Tokenizer: Learning Discrete Representations of Spatial Audio with Multichannel VQ-GAN
Host: Hannes Gamper, Microsoft Research Speaker: Parthasaarathy Sudarsanam, Tampere University Spatial audio captures the directional and environmental characteristics of sound, enabling immersive
260
180 days 56:10 Efficient Secure Aggregation for Federated Learning
Host: Melissa Chase, Microsoft Research Speaker: Varun Madathil, Yale University Federated Learning (FL) trains a global model by having each selected device push only its model update to a
412
10.2
182 days 1:16 What does it mean to paraphrase a protein?
This explainer video unpacks the science behind paraphrasing a protein—a defining technique of Microsoft’s Paraphrase Project.
742
6.1
190 days 1:07:22 Evaluating the Cultural Relevance of AI Models and Products: Insights from the YUX Team
How do we ensure that AI systems are not only technologically advanced but also deeply resonant and beneficial for the specific communities they aim to serve?
242
190 days 28:52 Evaluating Privacy Policies under Modern Privacy Laws At Scale: An LLM-Based Automated Approach
Website privacy policies detail an online service's information practices, including how they handle user data and rights.
448
7.9
191 day 52:07 Ideas: More AI-resilient biosecurity with the Paraphrase Project
Microsoft’s Eric Horvitz and guests Tessa Alexanian, James Diggans, and Bruce Wittmann discuss the Paraphrase Project—a red-teaming effort that exposed and secured a biosecurity vulnerability in
539
9.6
196 days 29:05 Detecting Compromise of Passkey Storage on the Cloud
FIDO synced passkeys address account recovery challenges by enabling users to back up their FIDO2 private signing keys to the cloud storage of passkey management services (PMS).
413
8.8
196 days 34:32 Encrypted Access Logging for Online Accounts: Device Attributions without Device Tracking
Despite improvements in authentication mechanisms, compromise of online accounts remains prevalent. Therefore, technologies to detect compromise retroactively are also necessary.
267
197 days 6:04 Microsoft Research India - The evolution
Learn about Microsoft Research India’s journey from its inception to becoming a leading research center in computer science in India. Learn more about Microsoft Research India: microsoft.
1 055
11.8
197 days 7:01 Microsoft Research India - The lab culture
Watch the founder, former and current leaders, and researchers at the Microsoft Research India lab talk about its unique culture.
11 737
12.7
203 days 29:56 Understanding How Users Prepare for and React to Smartphone Theft
Smartphone theft is common, yet little research explores how users prepare for or respond to such incidents.
551
18
203 days 30:54 When LLMs Go Online: The Emerging Threat of Web-Enabled LLMs
Recent advancements in Large Language Models (LLMs) have established them as agentic systems capable of planning and interacting with various tools.
489
9.2
203 days 31:59 A Formal Analysis of Apple’s iMessage PQ3 Protocol
We present the formal verification of Apple's iMessage PQ3, a highly performant, device-to-device messaging protocol offering strong security guarantees even against an adversary with quantum
426
9.1
203 days 31:34 Email Spoofing with SMTP Smuggling: How the Shared Email Infrastructures Magnify this Vulnerability
Email spoofing is a critical technique used in phishing attacks to impersonate a trusted sender.
375
9.2
203 days 33:23 A Framework for Abusability Analysis: The Case of Passkeys in Interpersonal Threat Models
The recent rollout of passkeys by hundreds of web services online is the largest attempt yet to achieve the goal of passwordless authentication.
184
203 days 28:32 ‘Hey mum, I dropped my phone down the toilet’: Investigating Hi Mum and Dad SMS Scams in the UK
SMS fraud has surged in recent years. Detection techniques have improved along with the fraud, necessitating harder-to-detect fraud techniques.
262
203 days 8:08 Dehumanizing machines: Making sense of AI systems that seem human
Alexandra Olteanu, Principal Researcher, and Su Lin Blodgett, Principal Researcher, both from Microsoft Research Montréal, discuss the anthropomorphic design and perception of generative AI systems.
1 186
14.3
203 days 9:24 Scalable emulation of protein equilibrium ensembles with BioEmu
Frank Noé, Partner Research Manager at Microsoft Research AI for Science, shares a major milestone in biomolecular simulation: BioEmu, an emulator that predicts protein shape changes and stabilities
405
10.9
203 days 9:57 Disrupting the AI infrastructure with MicroLEDs
Paolo Costa, Partner Research Manager at Microsoft Research Cambridge, presents a breakthrough in AI infrastructure: using microLEDs to overcome scaling limitations in current communication
1 419
12.4
203 days 11:16 Dion: The distributed orthonormal update revolution is here
Kwangjun Ahn, Senior Researcher at Microsoft Research AI Frontiers, introduces Dion, a next-generation optimizer in the style of Muon that orthonormalizes only the top-r subspace via amortized power
1 943
15.4
203 days 7:36 Pushing boundaries of complex reasoning in small language models
Mojan Javaheripi, Member of Technical Staff at Microsoft Research AI Frontiers, presents Phi-4-Reasoning and Phi-4-Reasoning-Plus—two 14B models designed to advance complex reasoning in small-scale
467
5.9
203 days 48:27 Microsoft Research Forum | Season 2, Episode 1
Microsoft Research Forum is a virtual series highlighting purposeful research and its real-world impact, from fundamental exploration to advancing AI responsibly, scaling innovation through products
1 741
12
205 days 23:37 zk-promises: Anonymous Moderation, Reputation, & Blocking from Anonymous Credentials with Callbacks
Anonymity is essential for free speech and expressing dissent, but platform moderators need ways to police bad actors.
363
11.8
205 days 32:50 More is Less: Extra Features in Contactless Payments Break Security
The EMV contactless payment system has many independent parties: payment providers, terminal companies, smartphone companies, banks and regulators.
465
5.4
209 days 53:42 Sub-Population Identification of Multi-morbidity in Sub-Saharan African Populations
This work bridges medical literature on multimorbidity with data science approaches to exploratory health data analysis.
239
223 days 9:21 Echoes in GenAI generations
In our recent PNAS paper we demonstrate that large language models produce little variation in generated narratives. Compared to those generations, a human-written narrative is usually Sui Generis, i.
444
10.9