no code implementations • 19 Jan 2024 • Prakhar Kaushik, Aayush Mishra, Adam Kortylewski, Alan Yuille
We focus on individual locally robust mesh vertex features and iteratively update them based on their proximity to corresponding features in the target domain even when the global pose is not correct.
Ranked #1 on Unsupervised Domain Adaptation on OOD-CV
no code implementations • 12 Oct 2023 • Lingfeng Shen, Aayush Mishra, Daniel Khashabi
Furthermore, the theoretical hand-constructed weights used in these studies have properties that don't match those of real LLMs.
no code implementations • 28 Sep 2023 • Aayush Mishra, Karan Thakkar
This report examines the effectiveness of Chain-of-Thought (CoT) prompting in improving the multi-step reasoning abilities of large language models (LLMs).
no code implementations • 6 Sep 2023 • Aditya Mishra, Haroon R. Lone, Aayush Mishra
In summary, our research contributes to energy prediction by offering a robust LSTM model that outperforms alternative methods and operates with remarkable efficiency, generalizability, and reliability.
1 code implementation • 26 Jul 2022 • Aayush Mishra, Anqi Liu
The EI step uses a reference model which focuses on spurious correlations to efficiently reach a good environment partition.
1 code implementation • ? 2019 • Aayush Mishra, Suraj Kumar, Aditya Nigam, Saiful Islam
Traditional image steganography techniques hide the secret image intohigh-frequency regions of the cover images.