English
Related papers

Related papers: DAISY: Data Adaptive Self-Supervised Early Exit fo…

200 papers

Large Reasoning Models (LRMs) often suffer from \emph{overthinking}, a phenomenon in which redundant reasoning steps are generated after a correct solution has already been reached. Existing early reasoning exit methods primarily rely on…

Computation and Language · Computer Science 2026-04-17 Kang Liu , Yongkang Liu , Xiaocui Yang , Peidong Wang , Wen Zhang , Shi Feng , Yifei Zhang , Daling Wang

Self-supervised learning (SSL) models reshaped our approach to speech, language and vision. However their huge size and the opaque relations between their layers and tasks result in slow inference and network overthinking, where predictions…

Computation and Language · Computer Science 2022-11-17 Dan Berrebbi , Brian Yan , Shinji Watanabe

Deep neural networks have become larger over the years with increasing demand of computational resources for inference; incurring exacerbate costs and leaving little room for deployment on devices with limited battery and other resources…

Machine Learning · Computer Science 2021-09-28 Aaqib Saeed

Deep neural networks (DNNs) have been successfully applied in various fields. In DNNs, a large number of multiply-accumulate (MAC) operations are required to be performed, posing critical challenges in applying them in resource-constrained…

Machine Learning · Computer Science 2024-02-20 Jingcun Wang , Bing Li , Grace Li Zhang

Early exiting allows instances to exit at different layers according to the estimation of difficulty. Previous works usually adopt heuristic metrics such as the entropy of internal outputs to measure instance difficulty, which suffers from…

Computation and Language · Computer Science 2022-03-04 Tianxiang Sun , Xiangyang Liu , Wei Zhu , Zhichao Geng , Lingling Wu , Yilong He , Yuan Ni , Guotong Xie , Xuanjing Huang , Xipeng Qiu

Currently, pre-trained models can be considered the default choice for a wide range of NLP tasks. Despite their SoTA results, there is practical evidence that these models may require a different number of computing layers for different…

Machine Learning · Computer Science 2023-05-19 Nikita Balagansky , Daniil Gavrilov

As Internet of Things (IoT) technology advances, end devices like sensors and smartphones are progressively equipped with AI models tailored to their local memory and computational constraints. Local inference reduces communication costs…

Machine Learning · Computer Science 2024-08-22 Caelin Kaplan , Angelo Rodio , Tareq Si Salem , Chuan Xu , Giovanni Neglia

Despite the great success in Natural Language Processing (NLP) area, large pre-trained language models like BERT are not well-suited for resource-constrained or real-time applications owing to the large number of parameters and slow…

Computation and Language · Computer Science 2021-07-02 Keli Xie , Siyuan Lu , Meiqi Wang , Zhongfeng Wang

Distributed DNN inference is becoming increasingly important as the demand for intelligent services at the network edge grows. By leveraging the power of distributed computing, edge devices can perform complicated and resource-hungry…

Distributed, Parallel, and Cluster Computing · Computer Science 2025-02-25 Xian Peng , Xin Wu , Lianming Xu , Li Wang , Aiguo Fei

We introduce a two-dimensional (2D) early exit strategy that coordinates layer-wise and sentence-wise exiting for classification tasks in large language models. By processing input incrementally sentence-by-sentence while progressively…

Computation and Language · Computer Science 2026-04-22 Jan Hůla , David Adamczyk , Tomáš Filip , Martin Pavlíček , Petr Sosík

Sustainable artificial intelligence focuses on data, hardware, and algorithms to make machine learning models more environmentally responsible. In particular, machine learning models for speech representations are computationally expensive,…

Computation and Language · Computer Science 2024-06-13 Luis Lugo , Valentin Vielzeuf

Early Exit (EE) techniques have emerged as a means to reduce inference latency in Deep Neural Networks (DNNs). The latency improvement and accuracy in these techniques crucially depend on the criteria used to make exit decisions. We propose…

Machine Learning · Computer Science 2025-02-04 Divya Jyoti Bajpai , Manjesh Kumar Hanawal

Early-Exit (EE) is a Large Language Model (LLM) architecture that accelerates inference by allowing easier tokens to be generated using only a subset of the model's layers. However, traditional batching frameworks are ill-suited for EE…

Distributed, Parallel, and Cluster Computing · Computer Science 2025-12-18 Xuting Liu , Daniel Alexander , Siva Kesava Reddy Kakarla , Behnaz Arzani , Vincent Liu

Deep Learning is becoming increasingly relevant in Embedded and Internet-of-things applications. However, deploying models on embedded devices poses a challenge due to their resource limitations. This can impact the model's inference…

Machine Learning · Computer Science 2024-03-14 Max Sponner , Lorenzo Servadei , Bernd Waschneck , Robert Wille , Akash Kumar

Early Exiting is one of the most popular methods to achieve efficient inference. Current early exiting methods adopt the (weighted) sum of the cross entropy loss of all internal classifiers during training, imposing all these classifiers to…

Computation and Language · Computer Science 2024-04-09 Ziqian Zeng , Yihuai Hong , Hongliang Dai , Huiping Zhuang , Cen Chen

To tackle the high inference latency exhibited by autoregressive language models, previous studies have proposed an early-exiting framework that allocates adaptive computation paths for each token based on the complexity of generating the…

Computation and Language · Computer Science 2023-10-10 Sangmin Bae , Jongwoo Ko , Hwanjun Song , Se-Young Yun

Early exiting is an effective paradigm for improving the inference efficiency of pre-trained language models (PLMs) by dynamically adjusting the number of executed layers for each sample. However, in most existing works, easy and hard…

Machine Learning · Computer Science 2024-12-19 Jianing He , Qi Zhang , Hongyun Zhang , Xuanjing Huang , Usman Naseem , Duoqian Miao

The pre-trained language models like BERT, though powerful in many natural language processing tasks, are both computation and memory expensive. To alleviate this problem, one approach is to compress them for specific tasks before…

Computation and Language · Computer Science 2020-10-12 Lu Hou , Zhiqi Huang , Lifeng Shang , Xin Jiang , Xiao Chen , Qun Liu

As NLP models become larger, executing a trained model requires significant computational resources incurring monetary and environmental costs. To better respect a given inference budget, we propose a modification to contextual…

Computation and Language · Computer Science 2020-05-12 Roy Schwartz , Gabriel Stanovsky , Swabha Swayamdipta , Jesse Dodge , Noah A. Smith

Encoder-decoder transformer models have achieved great success on various vision-language (VL) tasks, but they suffer from high inference latency. Typically, the decoder takes up most of the latency because of the auto-regressive decoding.…

Computer Vision and Pattern Recognition · Computer Science 2023-11-16 Peng Tang , Pengkai Zhu , Tian Li , Srikar Appalaraju , Vijay Mahadevan , R. Manmatha