At a glance AI benchmarks report performance on specific tasks but provide limited insight into underlying capabilities; ADeLe evaluates models by scoring both tasks and models across 18 core abilities, enabling direct comparison between task demands and model capabilities. Using these ability scores, the  

[MUSIC FADES]  I’d like to ask each of my guests to introduce themselves. Tell me a little bit about your background and what you’re currently working on—to the extent you can talk about it—in AI. So, Nicolò, would you please start?  NICOLÒ FUSI: Yeah, thank you, Doug, for having us and having me here. It’s so much fun. So I’m Nicolò Fusi. I’m a researcher at MSR [Microsoft Research]. So Doug is my boss, so I will be very, very, very good to Doug in this podcast.   No, but jokes aside, my own background is in Bayesian nonparametric. That’s what I started studying. So […]

At a glance Problem: Debugging AI agent failures is hard because trajectories are long, stochastic, and often multi-agent, so the true root cause gets buried. Solution: AgentRx (opens in new tab) pinpoints the first unrecoverable (“critical failure”) step by synthesizing guarded, executable constraints  

At a glance Today’s AI agents store long interaction histories but struggle to reuse them effectively. Raw memory retrieval can overwhelm agents with lengthy, low-value context. PlugMem transforms interaction history into structured, reusable knowledge. A single, general-purpose memory module improves performance across diverse agent benchmarks while using fewer memory tokens.  

Transcript [MUSIC]  DOUG BURGER: AI is going to reshape the future. I don’t think there’s any question about that now. How we reshape it depends on the choices we make, and so it’s important to understand what we think those shapes are.  This is The Shape of Things to Come. I’m Doug Burger. I manage Microsoft Research’s worldwide labs, and I’m excited to introduce this new Microsoft Research Podcast series.   I called the podcast The Shape of Things to Come because as researchers, the problems that we choose to solve and the technologies that we develop do […]

At a glance Today’s AI agent benchmarks test one task at a time, while real workplace productivity requires managing dozens of interdependent tasks at once. To reflect this, we created a setting called Multi-Horizon Task Environments (MHTEs). Under multi-task loads, leading computer-using agents degrade sharply, with completion rates dropping from 16.7% to 8.7%. CORPGEN introduces digital employees,  

At a glance Imitation learning becomes easier when an AI agent understands why an action is taken. Predictive Inverse Dynamics Models (PIDMs) predict plausible future states, clarifying the direction of behavior during imitation learning. Even imperfect predictions reduce ambiguity, making it clearer which action makes sense in the moment. This makes PIDMs far more data‑efficient  

At a glance AI-driven medical image report generation can help medical providers become more efficient and productive. Current models are difficult to train because reporting practices vary widely among providers. Universal Report Generation (UniRG) uses reinforcement learning to align model training with real-world radiology practice rather than proxy text-generation objectives. UniRG  

At a glance Today’s multimodal AI systems can give answers that sound right but may not be grounded in what they actually observe over time, leading to unpredictable errors and safety risks in real-world settings. Argos is a verification framework