Zobrazeno 1 - 10
of 2 761
pro vyhledávání: '"Jiwoo An"'
Disentangled representation learning (DRL) aims to break down observed data into core intrinsic factors for a profound understanding of the data. In real-world scenarios, manually defining and labeling these factors are non-trivial, making unsupervis
Externí odkaz:
http://arxiv.org/abs/2410.23820
Autor:
Ullrich, Paul A., Barnes, Elizabeth A., Collins, William D., Dagon, Katherine, Duan, Shiheng, Elms, Joshua, Lee, Jiwoo, Leung, L. Ruby, Lu, Dan, Molina, Maria J., O'Brien, Travis A.
Machine learning (ML) is a revolutionary technology with demonstrable applications across multiple disciplines. Within the Earth science community, ML has been most visible for weather forecasting, producing forecasts that rival modern physics-based
Externí odkaz:
http://arxiv.org/abs/2410.19882
Reinforcement learning with human feedback (RLHF) is shown to largely benefit from precise reward models (RMs). However, recent studies in reward modeling schemes are skewed towards English, limiting the applicability of RLHF in multilingual alignmen
Externí odkaz:
http://arxiv.org/abs/2410.18027
In abstractive summarization, the challenge of producing concise and accurate summaries arises from the vast amount of information contained in the source document. Consequently, although Large Language Models (LLMs) can generate fluent text, they of
Externí odkaz:
http://arxiv.org/abs/2409.18618
Stable pre-training is essential for achieving better-performing language models. However, tracking pre-training stability by calculating gradient variance at every step is impractical due to the significant computational costs. We explore Token Embe
Externí odkaz:
http://arxiv.org/abs/2409.07787
Autor:
Berto, Federico, Hua, Chuanbo, Luttmann, Laurin, Son, Jiwoo, Park, Junyoung, Ahn, Kyuree, Kwon, Changhyun, Xie, Lin, Park, Jinkyoo
Multi-agent combinatorial optimization problems such as routing and scheduling have great practical relevance but present challenges due to their NP-hard combinatorial nature, hard constraints on the number of possible agents, and hard-to-optimize ob
Externí odkaz:
http://arxiv.org/abs/2409.03811
Generative models have become widely used in biomedical entity linking (BioEL) due to their excellent performance and efficient memory usage. However, these models are usually trained only with positive samples--entities that match the input mention'
Externí odkaz:
http://arxiv.org/abs/2408.16493
Representing symbolic music with compound tokens, where each token consists of several different sub-tokens representing a distinct musical feature or attribute, offers the advantage of reducing sequence length. While previous research has validated
Externí odkaz:
http://arxiv.org/abs/2408.01180
Autor:
Jeong, Inho, Yang, Jiwoo, Jang, Juntae, Cho, Daeheum, Kwon, Deok-Hwang, Kim, Jae-Keun, Lee, Takhee, Cho, Kyungjune, Chung, Seungjun
In practical electronic applications, where doping is crucial to exploit large-area two-dimensional (2D) semiconductors, surface charge transfer doping (SCTD) has emerged as a promising strategy to tailor their electrical characteristics. However, im
Externí odkaz:
http://arxiv.org/abs/2407.21678
Crime situations are race against time. An AI-assisted criminal investigation system, providing prompt but precise legal counsel is in need for police officers. We introduce LAPIS (Language Model Augmented Police Investigation System), an automated s
Externí odkaz:
http://arxiv.org/abs/2407.20248