Harsha Noriの肖像

Harsha Nori

Director, Research Engineering

ニュース&特集

A visual illustration of Medprompt performance on the MedQA benchmark. Moving from left to right on a horizontal line, the illustration shows how different Medprompt components and additive contributions improve accuracy starting with zero-shot at 81.7 accuracy, to random few-shot at 83.9 accuracy, to random few-shot, chain-of-thought at 87.3 accuracy, to kNN, few-shot, chain-of-thought at 88.4 accuracy, to ensemble with choice shuffle at 90.2 accuracy.
Microsoft Research ブログ

| Eric Horvitz, Harsha Nori, と Naoto Usuyama

Discover the most effective run-time str…

three conversation bubbles on a blue, purple, and pink gradient background
Microsoft Research ブログ

| Eric Horvitz, Harsha Nori, と Yin Tat Lee

We’re seeing exciting capabilities of fr…

Microsoft Research ブログ

At the 2005 Conference on Neural Informa…

a man standing in front of a building
Microsoft Research ブログ

| Rich Caruana, Harsha Nori, Samuel Jenkins, Paul Koch, と Ester de Nicolas

When AI systems impact people’s lives, i…