English
Related papers

Related papers: Structural Sensitivity in Compressed Transformers:…

200 papers

Deep learning approaches have achieved unprecedented performance in visual recognition tasks such as object detection and pose estimation. However, state-of-the-art models have millions of parameters represented as floats which make them…

Computer Vision and Pattern Recognition · Computer Science 2021-02-08 Gedeon Muhawenayo , Georgia Gkioxari

Attention layers, as commonly used in transformers, form the backbone of modern deep learning, yet there is no mathematical description of their benefits and deficiencies as compared with other architectures. In this work we establish both…

Machine Learning · Computer Science 2023-11-17 Clayton Sanford , Daniel Hsu , Matus Telgarsky

Deep Neural Networks have been used in a wide variety of applications with significant success. However, their highly complex nature owing to comprising millions of parameters has lead to problems during deployment in pipelines with low…

Machine Learning · Computer Science 2022-08-15 Elvis Johnson , Xiaochen Tang , Sriramacharyulu Samudrala

The recent trend in industry-setting Natural Language Processing (NLP) research has been to operate large %scale pretrained language models like BERT under strict computational limits. While most model compression work has focused on…

Computation and Language · Computer Science 2021-04-13 J. S. McCarley , Rishav Chakravarti , Avirup Sil

Pruning the weights of neural networks is an effective and widely-used technique for reducing model size and inference complexity. We develop and test a novel method based on compressed sensing which combines the pruning and training into a…

Computer Vision and Pattern Recognition · Computer Science 2021-04-08 Jonathan W. Siegel , Jianhong Chen , Pengchuan Zhang , Jinchao Xu

How is knowledge stored in an LLM's weights? We study this via layer pruning: if removing a certain layer does not affect model performance in common question-answering benchmarks, then the weights in that layer are not necessary for…

Computation and Language · Computer Science 2025-03-04 Andrey Gromov , Kushal Tirumala , Hassan Shapourian , Paolo Glorioso , Daniel A. Roberts

Low-rank tensor compression has been proposed as a promising approach to reduce the memory and compute requirements of neural networks for their deployment on edge devices. Tensor compression reduces the number of parameters required to…

Machine Learning · Computer Science 2021-11-03 Cole Hawkins , Haichuan Yang , Meng Li , Liangzhen Lai , Vikas Chandra

We challenge the conventional view of neural network pruning as solely a compression technique, demonstrating that one-shot magnitude pruning serves as a powerful implicit regularizer for ASR. Using Whisper-small, we combine gradient- and…

Audio and Speech Processing · Electrical Eng. & Systems 2025-11-12 Julian Irigoyen , Arthur Söhler , Andreas Søeborg Kirkedal

Large Language Models (LLMs) demonstrate exceptional reasoning abilities, enabling strong generalization across diverse tasks such as commonsense reasoning and instruction following. However, as LLMs scale, inference costs become…

Computation and Language · Computer Science 2025-02-06 Rhea Sanjay Sukthanker , Benedikt Staffler , Frank Hutter , Aaron Klein

Why do language models trained on contradictory data prefer correct answers? In controlled experiments with small transformers (3.5M--86M parameters), we show that this preference tracks the compressibility structure of errors rather than…

Computation and Language · Computer Science 2026-04-07 Konstantin Krestnikov

The goal of model compression is to reduce the size of a large neural network while retaining a comparable performance. As a result, computation and memory costs in resource-limited applications may be significantly reduced by dropping…

Machine Learning · Statistics 2022-11-10 Wenjing Yang , Ganghua Wang , Jie Ding , Yuhong Yang

Transformer based large language models have achieved tremendous success. However, the significant memory and computational costs incurred during the inference process make it challenging to deploy large models on resource-constrained…

Computation and Language · Computer Science 2024-02-16 Wenxiao Wang , Wei Chen , Yicong Luo , Yongliu Long , Zhengkai Lin , Liye Zhang , Binbin Lin , Deng Cai , Xiaofei He

Weight pruning is a standard technique for compressing large language models, yet its effect on learned internal representations remains poorly understood. We present the first systematic study of how unstructured pruning reshapes the…

Machine Learning · Computer Science 2026-03-27 Hector Borobia , Elies Seguí-Mas , Guillermina Tormo-Carbó

Model compression and hardware acceleration are essential for the resource-efficient deployment of deep neural networks. Modern object detectors have highly interconnected convolutional layers with concatenations. In this work, we study how…

Computer Vision and Pattern Recognition · Computer Science 2024-05-08 Svetlana Pavlitska , Oliver Bagge , Federico Peccia , Toghrul Mammadov , J. Marius Zöllner

Although large language models (LLMs) have achieved remarkable success across various domains, their considerable scale necessitates substantial computational resources, posing significant challenges for deployment in resource-constrained…

Machine Learning · Computer Science 2024-11-26 Yao Lu , Hao Cheng , Yujie Fang , Zeyu Wang , Jiaheng Wei , Dongwei Xu , Qi Xuan , Xiaoniu Yang , Zhaowei Zhu

Top-down induction of decision trees has been observed to suffer from the inadequate functioning of the pruning phase. In particular, it is known that the size of the resulting tree grows linearly with the sample size, even though the…

Artificial Intelligence · Computer Science 2011-06-06 T. Elomaa , M. Kaariainen

With the wide adoption of language models for IR -- and specifically RAG systems -- the latency of the underlying LLM becomes a crucial bottleneck, since the long contexts of retrieved passages lead large prompts and therefore, compute…

Information Retrieval · Computer Science 2026-04-06 Cornelius Kummer , Lena Jurkschat , Michael Färber , Sahar Vahdati

This micro-paper describes a trick to speed up inference of transformers with RoPE (such as LLaMA, Mistral, PaLM, and Gemma). For these models, a large portion of the first transformer layer can be precomputed, which results in slightly…

Machine Learning · Computer Science 2024-03-13 Nils Graef

In this paper, we propose a rotation-constrained compensation method to address the errors introduced by structured pruning of large language models (LLMs). LLMs are trained on massive datasets and accumulate rich semantic knowledge in…

Computation and Language · Computer Science 2026-03-02 Shuichiro Haruta , Kazunori Matsumoto , Zhi Li , Yanan Wang , Mori Kurokawa

Reasoning language models such as DeepSeek-R1 produce long chain-of-thought traces during inference time which make them costly to deploy at scale. We show that using compression techniques such as neural network pruning produces greater…

Artificial Intelligence · Computer Science 2026-05-05 Ryan Lucas , Kayhan Behdin , Zhipeng Wang , Qingquan Song , Shao Tang , Rahul Mazumder