English
Related papers

Related papers: Your representations are in the network: composabl…

200 papers

Despite the success of fine-tuning pretrained language encoders like BERT for downstream natural language understanding (NLU) tasks, it is still poorly understood how neural networks change after fine-tuning. In this work, we use centered…

Computation and Language · Computer Science 2021-09-21 Jason Phang , Haokun Liu , Samuel R. Bowman

When adapting large language models (LLMs) to a specific downstream task, two primary approaches are commonly employed: (1) prompt engineering, often with in-context few-shot learning, leveraging the model's inherent generalization…

Machine Learning · Computer Science 2025-12-24 Jorg Bornschein , Clare Lyle , Yazhe Li , Amal Rannen-Triki , Xu Owen He , Razvan Pascanu

Transformer-based pre-trained models, such as BERT, have shown extraordinary success in achieving state-of-the-art results in many natural language processing applications. However, deploying these models can be prohibitively costly, as the…

Computation and Language · Computer Science 2022-03-18 Jing Zhao , Yifan Wang , Junwei Bao , Youzheng Wu , Xiaodong He

Recent studies have explored the working mechanisms of In-Context Learning (ICL). However, they mainly focus on classification and simple generation tasks, limiting their broader application to more complex generation tasks in practice. To…

Computation and Language · Computer Science 2025-03-14 Zhenyu Liu , Dongfang Li , Xinshuo Hu , Xinping Zhao , Yibin Chen , Baotian Hu , Min Zhang

Conventional wisdom suggests that pre-training Vision Transformers (ViT) improves downstream performance by learning useful representations. Is this actually true? We investigate this question and find that the features and representations…

Machine Learning · Computer Science 2024-11-15 Alexander C. Li , Yuandong Tian , Beidi Chen , Deepak Pathak , Xinlei Chen

The attention mechanism is considered the backbone of the widely-used Transformer architecture. It contextualizes the input by computing input-specific attention matrices. We find that this mechanism, while powerful and elegant, is not as…

Computation and Language · Computer Science 2022-11-08 Michael Hassid , Hao Peng , Daniel Rotem , Jungo Kasai , Ivan Montero , Noah A. Smith , Roy Schwartz

This paper proposes Skip2-LoRA as a lightweight fine-tuning method for deep neural networks to address the gap between pre-trained and deployed models. In our approach, trainable LoRA (low-rank adaptation) adapters are inserted between the…

Machine Learning · Computer Science 2024-10-29 Hiroki Matsutani , Masaaki Kondo , Kazuki Sunaga , Radu Marculescu

We present FIT: a transformer-based architecture with efficient self-attention and adaptive computation. Unlike original transformers, which operate on a single sequence of data tokens, we divide the data tokens into groups, with each group…

Machine Learning · Computer Science 2023-05-26 Ting Chen , Lala Li

Large pre-trained speech models are widely used as the de-facto paradigm, especially in scenarios when there is a limited amount of labeled data available. However, finetuning all parameters from the self-supervised learned model can be…

Audio and Speech Processing · Electrical Eng. & Systems 2023-06-16 Nanxin Chen , Izhak Shafran , Yu Zhang , Chung-Cheng Chiu , Hagen Soltau , James Qin , Yonghui Wu

Parameter efficient adaptation methods have become a key mechanism to train large pre-trained models for downstream tasks. However, their per-task parameter overhead is considered still high when the number of downstream tasks to adapt for…

Audio and Speech Processing · Electrical Eng. & Systems 2024-04-01 Tsendsuren Munkhdalai , Youzheng Chen , Khe Chai Sim , Fadi Biadsy , Tara Sainath , Pedro Moreno Mengibar

Foundation models have revolutionized AI, but adapting them efficiently for multimodal tasks, particularly in dual-stream architectures composed of unimodal encoders, such as DINO and BERT, remains a significant challenge.…

Computer Vision and Pattern Recognition · Computer Science 2026-04-07 Wish Suharitdamrong , Tony Alex , Muhammad Awais , Sara Ahmed

Attention calculation is extremely time-consuming for long-sequence inference tasks, such as text or image/video generation, in large models. To accelerate this process, we developed a low-precision, mathematically-equivalent algorithm…

Fine-tuning provides an effective means to specialize pre-trained models for various downstream tasks. However, fine-tuning often incurs high memory overhead, especially for large transformer-based models, such as LLMs. While existing…

Computation and Language · Computer Science 2025-02-03 Antoine Simoulin , Namyong Park , Xiaoyi Liu , Grey Yang

Cross Attention is a popular method for retrieving information from a set of context tokens for making predictions. At inference time, for each prediction, Cross Attention scans the full set of $\mathcal{O}(N)$ tokens. In practice, however,…

Machine Learning · Computer Science 2024-03-04 Leo Feng , Frederick Tung , Hossein Hajimirsadeghi , Yoshua Bengio , Mohamed Osama Ahmed

The goal of continual learning is to improve the performance of recognition models in learning sequentially arrived data. Although most existing works are established on the premise of learning from scratch, growing efforts have been…

Computer Vision and Pattern Recognition · Computer Science 2023-10-10 Gengwei Zhang , Liyuan Wang , Guoliang Kang , Ling Chen , Yunchao Wei

Foundation models have achieved great advances in multi-task learning with a unified interface of unimodal and multimodal tasks. However, the potential of such multi-task learners has not been exploited during transfer learning. In this…

Computer Vision and Pattern Recognition · Computer Science 2023-05-18 Chengyue Wu , Teng Wang , Yixiao Ge , Zeyu Lu , Ruisong Zhou , Ying Shan , Ping Luo

Existing adaptation techniques typically require architectural modifications or added parameters, leading to high computational costs and complexity. We introduce Attention Projection Layer Adaptation (APLA), a simple approach to adapt…

Computer Vision and Pattern Recognition · Computer Science 2025-03-25 Moein Sorkhei , Emir Konuk , Kevin Smith , Christos Matsoukas

Large pre-trained models are commonly adapted to downstream tasks using parameter-efficient fine-tuning methods such as Low-Rank Adaptation (LoRA), which injects small trainable low-rank matrices instead of updating all weights. While LoRA…

Machine Learning · Computer Science 2026-03-10 Nurbek Tastan , Stefanos Laskaridis , Martin Takac , Karthik Nandakumar , Samuel Horvath

Finetuning a pretrained model has become a standard approach for training neural networks on novel tasks, resulting in fast convergence and improved performance. In this work, we study an alternative finetuning method, where instead of…

Machine Learning · Computer Science 2023-07-04 Gal Kaplun , Andrey Gurevich , Tal Swisa , Mazor David , Shai Shalev-Shwartz , Eran Malach

Causal effect estimation from observational data requires careful adjustment for confounding. Classical estimators such as inverse probability weighting and augmented inverse probability weighting are effective under favorable model…

Machine Learning · Statistics 2026-04-28 Lei Wang , Debashis Ghosh