Zobrazeno 1 - 10
of 394
pro vyhledávání: '"Chen, Guangyi"'
Autor:
Liu, Zuyan, Liu, Benlin, Wang, Jiahui, Dong, Yuhao, Chen, Guangyi, Rao, Yongming, Krishna, Ranjay, Lu, Jiwen
In the field of instruction-following large vision-language models (LVLMs), the efficient deployment of these models faces challenges, notably due to the high memory demands of their key-value (KV) caches. Conventional cache management strategies for
Externí odkaz:
http://arxiv.org/abs/2407.18121
Recent works have introduced GNN-to-MLP knowledge distillation (KD) frameworks to combine both GNN's superior performance and MLP's fast inference speed. However, existing KD frameworks are primarily designed for node classification within single gra
Externí odkaz:
http://arxiv.org/abs/2406.19832
To better interpret the intrinsic mechanism of large language models (LLMs), recent studies focus on monosemanticity on its basic units. A monosemantic neuron is dedicated to a single and specific concept, which forms a one-to-one correlation between
Externí odkaz:
http://arxiv.org/abs/2406.17969
Learning concepts from natural high-dimensional data (e.g., images) holds potential in building human-aligned and interpretable machine learning models. Despite its encouraging prospect, formalization and theoretical insights into this crucial task a
Externí odkaz:
http://arxiv.org/abs/2406.00519
Autor:
Cai, Ruichu, Jiang, Zhifang, Li, Zijian, Chen, Weilin, Chen, Xuexin, Hao, Zhifeng, Shen, Yifan, Chen, Guangyi, Zhang, Kun
Existing methods for multi-modal time series representation learning aim to disentangle the modality-shared and modality-specific latent variables. Although achieving notable performances on downstream tasks, they usually assume an orthogonal latent
Externí odkaz:
http://arxiv.org/abs/2405.16083
Autor:
Li, Zijian, Shen, Yifan, Zheng, Kaitao, Cai, Ruichu, Song, Xiangchen, Gong, Mingming, Zhu, Zhengmao, Chen, Guangyi, Zhang, Kun
Temporally causal representation learning aims to identify the latent causal process from time series observations, but most methods require the assumption that the latent causal processes do not have instantaneous relations. Although some recent met
Externí odkaz:
http://arxiv.org/abs/2405.15325
In this paper, we investigate a new problem called narrative action evaluation (NAE). NAE aims to generate professional commentary that evaluates the execution of an action. Unlike traditional tasks such as score-based action quality assessment and v
Externí odkaz:
http://arxiv.org/abs/2404.14471
Autor:
Li, Zijian, Cai, Ruichu, Yang, Zhenhui, Huang, Haiqin, Chen, Guangyi, Shen, Yifan, Chen, Zhengming, Song, Xiangchen, Zhang, Kun
Temporal distribution shifts are ubiquitous in time series data. One of the most popular methods assumes that the temporal distribution shift occurs uniformly to disentangle the stationary and nonstationary dependencies. But this assumption is diffic
Externí odkaz:
http://arxiv.org/abs/2402.12767
Autor:
Li, Loka, Ng, Ignavier, Luo, Gongxu, Huang, Biwei, Chen, Guangyi, Liu, Tongliang, Gu, Bin, Zhang, Kun
Conventional causal discovery methods rely on centralized data, which is inconsistent with the decentralized nature of data in many real-world situations. This discrepancy has motivated the development of federated causal discovery (FCD) approaches.
Externí odkaz:
http://arxiv.org/abs/2402.13241
Few-shot action recognition aims at quickly adapting a pre-trained model to the novel data with a distribution shift using only a limited number of samples. Key challenges include how to identify and leverage the transferable knowledge learned by the
Externí odkaz:
http://arxiv.org/abs/2402.12706