1 code implementation • 21 Oct 2024 • Zhenpeng Su, Xing Wu, Zijia Lin, Yizhe Xiong, Minxuan Lv, Guangyuan Ma, Hui Chen, Songlin Hu, Guiguang Ding
Large language models (LLM) have been attracting much attention from the community recently, due to their remarkable performance in all kinds of downstream tasks.
no code implementations • 29 Sep 2024 • Xiao Wang, Jianlong Wu, Zijia Lin, Fuzheng Zhang, Di Zhang, Liqiang Nie
For iterative refinement, we first leverage a video-language model to generate synthetic annotations, resulting in a refined dataset.
1 code implementation • 10 Sep 2024 • Hui-Yue Yang, Hui Chen, Lihao Liu, Zijia Lin, Kai Chen, Liejun Wang, Jungong Han, Guiguang Ding
By incorporating the RASFormer block, our RAS method achieves superior contextual awareness capabilities, leading to remarkable performance.
no code implementations • 26 Jul 2024 • Mengyao Lyu, Tianxiang Hao, Xinhao Xu, Hui Chen, Zijia Lin, Jungong Han, Guiguang Ding
In response, we present learn from the learnt (LFTL), a novel paradigm for SFADA to leverage the learnt knowledge from the source pretrained model and actively iterated models without extra overhead.
1 code implementation • 13 Jul 2024 • Zhenpeng Su, Zijia Lin, Xue Bai, Xing Wu, Yizhe Xiong, Haoran Lian, Guangyuan Ma, Hui Chen, Guiguang Ding, Wei Zhou, Songlin Hu
Currently, the mainstream routing methods include dynamic routing and fixed routing.
no code implementations • 24 May 2024 • Chenxi Sun, Hongzhi Zhang, Zijia Lin, Jingyuan Zhang, Fuzheng Zhang, Zhongyuan Wang, Bin Chen, Chengru Song, Di Zhang, Kun Gai, Deyi Xiong
The core of our approach is the observation that a pre-trained language model can confidently predict multiple contiguous tokens, forming the basis for a \textit{lexical unit}, in which these contiguous tokens could be decoded in parallel.
3 code implementations • 23 May 2024 • Ao Wang, Hui Chen, Lihao Liu, Kai Chen, Zijia Lin, Jungong Han, Guiguang Ding
In this work, we aim to further advance the performance-efficiency boundary of YOLOs from both the post-processing and model architecture.
Ranked #18 on Real-Time Object Detection on MS COCO
1 code implementation • 29 Apr 2024 • Meng Li, Haoran Jin, Ruixuan Huang, Zhihao Xu, Defu Lian, Zijia Lin, Di Zhang, Xiting Wang
Based on this, we quantify the faithfulness of a concept explanation via perturbation.
no code implementations • 27 Apr 2024 • Haoran Lian, Yizhe Xiong, Jianwei Niu, Shasha Mo, Zhenpeng Su, Zijia Lin, Peng Liu, Hui Chen, Guiguang Ding
Due to their infrequent appearance in the text corpus, Scaffold Tokens pose a learning imbalance issue for language models.
no code implementations • 27 Apr 2024 • Yizhe Xiong, Xiansheng Chen, Xin Ye, Hui Chen, Zijia Lin, Haoran Lian, Zhenpeng Su, Jianwei Niu, Guiguang Ding
In this paper, we propose the novel concept of Temporal Scaling Law, studying how the test loss of an LLM evolves as the training steps scale up.
1 code implementation • 14 Mar 2024 • Yizhe Xiong, Hui Chen, Tianxiang Hao, Zijia Lin, Jungong Han, Yuesong Zhang, Guoxin Wang, Yongjun Bao, Guiguang Ding
Consequently, a simple combination of them cannot guarantee accomplishing both training efficiency and inference efficiency with minimal costs.
3 code implementations • 20 Jan 2024 • Guangyuan Ma, Xing Wu, Zijia Lin, Songlin Hu
In this study, we aim to shed light on this issue by revealing that masked auto-encoder (MAE) pre-training with enhanced decoding significantly improves the term coverage of input tokens in dense representations, compared to vanilla BERT checkpoints.
2 code implementations • 10 Dec 2023 • Ao Wang, Hui Chen, Zijia Lin, Jungong Han, Guiguang Ding
Here, to achieve real-time segmenting anything on mobile devices, following MobileSAM, we replace the heavyweight image encoder in SAM with RepViT model, ending up with the RepViT-SAM model.
1 code implementation • 30 Oct 2023 • Zhenpeng Su, Xing Wu, Xue Bai, Zijia Lin, Hui Chen, Guiguang Ding, Wei Zhou, Songlin Hu
Experiments reveal that models incorporating the proposed MiLe Loss can gain consistent performance improvement on downstream benchmarks.
Ranked #97 on Multi-task Language Understanding on MMLU
no code implementations • 11 Oct 2023 • Jiayi Fu, Lei Lin, Xiaoyang Gao, Pengli Liu, Zhengzong Chen, Zhirui Yang, ShengNan Zhang, Xue Zheng, Yan Li, Yuliang Liu, Xucheng Ye, Yiqiao Liao, Chao Liao, Bin Chen, Chengru Song, Junchen Wan, Zijia Lin, Fuzheng Zhang, Zhongyuan Wang, Di Zhang, Kun Gai
Recent advancements in large language models (LLMs) have demonstrated remarkable abilities in handling a variety of natural language processing (NLP) downstream tasks, even on mathematical tasks requiring multi-step reasoning.
Ranked #93 on Arithmetic Reasoning on GSM8K (using extra training data)
1 code implementation • ICCV 2023 • Yizhe Xiong, Hui Chen, Zijia Lin, Sicheng Zhao, Guiguang Ding
To address this issue, recent works consider the Few-shot Unsupervised Domain Adaptation (FUDA) where only a few source samples are labeled, and conduct knowledge transfer via self-supervised learning methods.
no code implementations • 27 Sep 2023 • Ao Wang, Hui Chen, Zijia Lin, Sicheng Zhao, Jungong Han, Guiguang Ding
We further employ a consistent dynamic channel pruning (CDCP) strategy to dynamically prune unimportant channels in ViTs.
no code implementations • 16 Aug 2023 • Guangyuan Ma, Xing Wu, Peng Wang, Zijia Lin, Songlin Hu
Concretely, we leverage the capabilities of LLMs for document expansion, i. e. query generation, and effectively transfer expanded knowledge to retrievers using pre-training strategies tailored for passage retrieval.
7 code implementations • CVPR 2024 • Ao Wang, Hui Chen, Zijia Lin, Jungong Han, Guiguang Ding
Recently, lightweight Vision Transformers (ViTs) demonstrate superior performance and lower latency, compared with lightweight Convolutional Neural Networks (CNNs), on resource-constrained mobile devices.
no code implementations • 5 Apr 2023 • Xing Wu, Guangyuan Ma, Peng Wang, Meng Lin, Zijia Lin, Fuzheng Zhang, Songlin Hu
As an effective representation bottleneck pretraining technique, the contextual masked auto-encoder utilizes contextual embedding to assist in the reconstruction of passages.
2 code implementations • 19 Dec 2022 • Xing Wu, Guangyuan Ma, Wanhui Qian, Zijia Lin, Songlin Hu
Recently, methods have been developed to improve the performance of dense passage retrieval by using context-supervised pre-training.
1 code implementation • 13 Oct 2022 • Xing Wu, Chaochen Gao, Zijia Lin, Zhongyuan Wang, Jizhong Han, Songlin Hu
Sparse sampling is also likely to miss important frames corresponding to some text portions, resulting in textual redundancy.
2 code implementations • 8 Oct 2022 • Xing Wu, Chaochen Gao, Zijia Lin, Jizhong Han, Zhongyuan Wang, Songlin Hu
Contrastive learning has been extensively studied in sentence embedding learning, which assumes that the embeddings of different views of the same sentence are closer.
2 code implementations • 16 Aug 2022 • Xing Wu, Guangyuan Ma, Meng Lin, Zijia Lin, Zhongyuan Wang, Songlin Hu
Dense passage retrieval aims to retrieve the relevant passages of a query from a large corpus based on dense representations (i. e., vectors) of the query and the passages.
1 code implementation • 15 Jul 2020 • Qianhui Wu, Zijia Lin, Börje F. Karlsson, Biqing Huang, Jian-Guang Lou
Prior works in cross-lingual named entity recognition (NER) with no/little labeled data fall into two primary categories: model transfer based and data transfer based methods.
Ranked #1 on Cross-Lingual NER on NoDaLiDa Norwegian Bokmål
no code implementations • 17 Jun 2020 • Yunqi Miao, Zijia Lin, Guiguang Ding, Jungong Han
In this paper, we propose a Shallow feature based Dense Attention Network (SDANet) for crowd counting from still images, which diminishes the impact of backgrounds via involving a shallow feature based attention model, and meanwhile, captures multi-scale information via densely connecting hierarchical image features.
1 code implementation • ACL 2020 • Qianhui Wu, Zijia Lin, Börje F. Karlsson, Jian-Guang Lou, Biqing Huang
However, such methods either are not applicable if the labeled data in the source languages is unavailable, or do not leverage information contained in unlabeled data in the target language.
Ranked #1 on Cross-Lingual NER on CoNLL German
1 code implementation • CVPR 2020 • Hui Chen, Guiguang Ding, Xudong Liu, Zijia Lin, Ji Liu, Jungong Han
Existing methods leverage the attention mechanism to explore such correspondence in a fine-grained manner.
Ranked #20 on Cross-Modal Retrieval on Flickr30k
1 code implementation • 14 Nov 2019 • Qianhui Wu, Zijia Lin, Guoxin Wang, Hui Chen, Börje F. Karlsson, Biqing Huang, Chin-Yew Lin
For languages with no annotated resources, transferring knowledge from rich-resource languages is an effective solution for named entity recognition (NER).
Ranked #1 on Cross-Lingual NER on MSRA
1 code implementation • 12 Jul 2019 • Hui Chen, Zijia Lin, Guiguang Ding, JianGuang Lou, Yusen Zhang, Borje Karlsson
The dominant approaches for named entity recognition (NER) mostly adopt complex recurrent neural networks (RNN), e. g., long-short-term-memory (LSTM).
Ranked #24 on Named Entity Recognition (NER) on Ontonotes v5 (English)
no code implementations • CVPR 2015 • Zijia Lin, Guiguang Ding, Mingqing Hu, Jian-Min Wang
With benefits of low storage costs and high query speeds, hashing methods are widely researched for efficiently retrieving large-scale data, which commonly contains multiple views, e. g. a news report with images, videos and texts.
no code implementations • CVPR 2013 • Zijia Lin, Guiguang Ding, Mingqing Hu, Jian-Min Wang, Xiaojun Ye
Though widely utilized for facilitating image management, user-provided image tags are usually incomplete and insufficient to describe the whole semantic content of corresponding images, resulting in performance degradations in tag-dependent applications and thus necessitating effective tag completion methods.