[Big Model Daily, December 16-17] Industry condolences! Tang Xiaoou, founder of SenseTime, passed away at the age of 55

Condolences! Tang Xiaoou, founder of SenseTime, passed away at the age of 55

https://news.miracleplus.com/share_link/13482

According to multiple sources, Tang Xiaoou, founder of SenseTime and professor at the Chinese University of Hong Kong, unfortunately passed away in his sleep at the age of 55. Tang Xiaoou graduated from the University of Science and Technology of China in 1990 and received his PhD from the Massachusetts Institute of Technology in 1996. Later, he entered the Chinese University of Hong Kong and founded the Multimedia Laboratory of the Chinese University of Hong Kong. He also worked in Microsoft Research Asia and other units. Tang Xiaoou has profound attainments in the field of computer vision, especially in the segmented track of face recognition, and he is a “pathfinder”.

Terence Tao’s favored proof assistant, Lean, uses large models

https://news.miracleplus.com/share_link/13483

The open platform LeanDojo team and researchers from the California Institute of Technology have launched Lean Copilot, a collaboration tool designed for LLM and human interaction, aiming to give 100% accurate formal mathematical proofs through human-machine collaboration.

The paper was rejected by the first ICLR and the code was over-optimized. Tomas Mikolov, the author of word2vec, shares the story behind it

https://news.miracleplus.com/share_link/13484

A few days ago, NeurIPS 2023 announced the winning papers, among which the Time Test Award was awarded to the NeurIPS paper “Distributed Representations of Words and Phrases and their Compositionality” ten years ago. This work introduced the groundbreaking word embedding technology word2vec, demonstrated the ability to learn from large amounts of unstructured text, and promoted the arrival of a new era of natural language processing. This paper was written by Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg Corrado, Jeffrey Dean and others who were all still at Google at the time. It has been cited more than 40,000 times.