by RLE at MIT | Feb 16, 2023 | News Links
The method enables a model to determine its confidence in a prediction, while using no additional data and far fewer computing resources than other methods. «more» Related Links: Efficient technique improves machine-learning models’ reliability (MIT News)...
by RLE at MIT | Feb 16, 2023 | RLE Recent Papers
Maohao Shen, Yuheng Bu, Prasanna Sattigeri, Soumya Ghosh, Subhro Das, Gregory Wornell doi: arXiv:2212.07359v1 Abstract: It is known that neural networks have the problem of being over-confident when directly using the output label distribution to generate...
by RLE at MIT | Jul 27, 2022 | RLE News Articles
Gary Lee’s research is focused on one of the dominant and enduring challenges in the field of communication: managing and rejecting interference in communication systems and networks, which otherwise strongly limit performance. The traditional approach of...
by RLE at MIT | Sep 20, 2019 | Awards, RLE News Articles
The Research Laboratory of Electronics (RLE) is pleased to announce that the 2019–2020 Claude E. Shannon Research Assistantship has been awarded to Anny Xijia Zheng and Ganesh Ajjaganadde, doctoral students in the Department of Electrical Engineering and...
by RLE at MIT | Jan 17, 2019 | RLE Recent Papers
Shao-Lun Huang, Gregory W. Wornell and Lizhong Zheng DOI: 10.1109/ITW.2018.8613343 Abstract: We address the problem of optimal feature selection for a Gaussian vector pair in the weak dependence regime, when the inference task is not known in advance....
