Zobrazeno 1 - 10
of 51
pro vyhledávání: '"Qin, Guanghui"'
Autor:
Hou, Abe Bohan, Weller, Orion, Qin, Guanghui, Yang, Eugene, Lawrie, Dawn, Holzenberger, Nils, Blair-Stanek, Andrew, Van Durme, Benjamin
Legal professionals need to write analyses that rely on citations to relevant precedents, i.e., previous case decisions. Intelligent systems assisting legal professionals in writing such documents provide great benefits but are challenging to design.
Externí odkaz:
http://arxiv.org/abs/2406.17186
Autor:
Rosset, Corby, Chung, Ho-Lam, Qin, Guanghui, Chau, Ethan C., Feng, Zhuo, Awadallah, Ahmed, Neville, Jennifer, Rao, Nikhil
Existing question answering (QA) datasets are no longer challenging to most powerful Large Language Models (LLMs). Traditional QA benchmarks like TriviaQA, NaturalQuestions, ELI5 and HotpotQA mainly study ``known unknowns'' with clear indications of
Externí odkaz:
http://arxiv.org/abs/2402.17896
Autor:
Tan, Weiting, Chen, Yunmo, Chen, Tongfei, Qin, Guanghui, Xu, Haoran, Zhang, Heidi C., Van Durme, Benjamin, Koehn, Philipp
We introduce STAR (Stream Transduction with Anchor Representations), a novel Transformer-based model designed for efficient sequence-to-sequence transduction over streams. STAR dynamically segments input streams to create compressed anchor representa
Externí odkaz:
http://arxiv.org/abs/2402.01172
Transformer-based language models (LMs) are inefficient in long contexts. We propose Dodo, a solution for context compression. Instead of one vector per token in a standard transformer model, Dodo represents text with a dynamic number of hidden state
Externí odkaz:
http://arxiv.org/abs/2310.02409
Autor:
Qin, Guanghui, Van Durme, Benjamin
Publikováno v:
ICML 2023
Embedding text sequences is a widespread requirement in modern language understanding. Existing approaches focus largely on constant-size representations. This is problematic, as the amount of information contained in text often varies with the lengt
Externí odkaz:
http://arxiv.org/abs/2310.01732
Publikováno v:
EACL 2023
Transformer models cannot easily scale to long sequences due to their O(N^2) time and space complexity. This has led to Transformer variants seeking to lower computational complexity, such as Longformer and Performer. While such models have theoretic
Externí odkaz:
http://arxiv.org/abs/2202.07856
Autor:
Yarmohammadi, Mahsa, Wu, Shijie, Marone, Marc, Xu, Haoran, Ebner, Seth, Qin, Guanghui, Chen, Yunmo, Guo, Jialiang, Harman, Craig, Murray, Kenton, White, Aaron Steven, Dredze, Mark, Van Durme, Benjamin
Zero-shot cross-lingual information extraction (IE) describes the construction of an IE model for some target language, given existing annotations exclusively in some other language, typically English. While the advance of pretrained multilingual enc
Externí odkaz:
http://arxiv.org/abs/2109.06798
Autor:
Qin, Guanghui, Eisner, Jason
Natural-language prompts have recently been used to coax pretrained language models into performing other AI tasks, using a fill-in-the-blank paradigm (Petroni et al., 2019) or a few-shot extrapolation paradigm (Brown et al., 2020). For example, lang
Externí odkaz:
http://arxiv.org/abs/2104.06599
Autor:
Xia, Patrick, Qin, Guanghui, Vashishtha, Siddharth, Chen, Yunmo, Chen, Tongfei, May, Chandler, Harman, Craig, Rawlins, Kyle, White, Aaron Steven, Van Durme, Benjamin
We present LOME, a system for performing multilingual information extraction. Given a text document as input, our core system identifies spans of textual entity and event mentions with a FrameNet (Baker et al., 1998) parser. It subsequently performs
Externí odkaz:
http://arxiv.org/abs/2101.12175
Copy mechanisms are employed in sequence to sequence models (seq2seq) to generate reproductions of words from the input to the output. These frameworks, operating at the lexical type level, fail to provide an explicit alignment that records where eac
Externí odkaz:
http://arxiv.org/abs/2010.15266