News & features
AI Frontiers: Rethinking intelligence with Ashley Llorens and Ida Momennejad
| Ashley Llorens and Ida Momennejad
Principal Researcher Ida Momennejad brings her expertise in cognitive neuroscience and computer science to this in-depth conversation about general intelligence and what the evolution of the brain across species can teach us about building AI.
Abstracts: January 25, 2024
| Gretchen Huizinga, Jordan Ash, and Dipendra Misra
On “Abstracts,” Jordan Ash & Dipendra Misra discuss the parameter reduction method LASER. Tune in to learn how selective removal of stored data alone can boost LLM performance, then sign up for Microsoft Research Forum for more on LASER &…
NeurIPS 2023 highlights breadth of Microsoft’s machine learning innovation
We’re proud to have 100+ accepted papers At NeurIPS 2023, plus 18 workshops. Several submissions were chosen as oral presentations and spotlight posters, reflecting groundbreaking concepts, methods, or applications. Here’s an overview of those submissions.
AI Frontiers: Measuring and mitigating harms with Hanna Wallach
| Hanna Wallach and Ashley Llorens
Powerful large-scale AI models like GPT-4 are showing dramatic improvements in reasoning, problem-solving, and language capabilities. This marks a phase change for artificial intelligence—and a signal of accelerating progress to come. In this Microsoft Research Podcast series, AI scientist and…
Inferring rewards through interaction
| Jessica Maghakian, Akanksha Saran, Cheng Tan, and Paul Mineiro
In reinforcement learning, handcrafting reward functions is difficult and can yield algorithms that don’t generalize well. IGL-P, an interaction-grounded learning strategy, learns personalized rewards for different people in recommender system scenarios.
In the news | Machine Learning (Theory)
HOMER: Provable Exploration in Reinforcement Learning
Last week at ICML 2020, Mikael Henaff, Akshay Krishnamurthy, John Langford and Dipendra Misra had a paper on a new reinforcement learning (RL) algorithm that solves three key problems in RL: (i) global exploration, (ii) decoding latent dynamics, and (iii) optimizing a given…
In the news | Medium | Machine Learning
HOMER: Provable Exploration in Reinforcement Learning
At ICML 2020, Mikael Henaff, Akshay Krishnamurthy, John Langford and Dipendra Misra published a paper presenting a new reinforcement learning (RL) algorithm called HOMER that addresses three main problems in real-world RL problem: (i) exploration, (ii) decoding latent dynamics, and (iii) optimizing…
Provably efficient reinforcement learning with Dr. Akshay Krishnamurthy
MSR’s New York City lab is home to some of the best reinforcement learning research on the planet but if you ask any of the researchers, they’ll tell you they’re very interested in getting it out of the lab and…
Provably efficient reinforcement learning with rich observations
| Akshay Krishnamurthy
Reinforcement learning, a machine learning paradigm for sequential decision making, has stormed into the limelight, receiving tremendous attention from both researchers and practitioners. When combined with deep learning, reinforcement learning (RL) has produced impressive empirical results, but the successes to…