Early-Exit Neural Networks

Follow this topic Ask about this topic

Momentum

4 papers in the last four weeks, against 1 the four weeks before. 0.0% of all new papers.

Jul 13Week of Sep 28

Latest papers 31

All topics

DEX: Digit-Level Early Exit for Energy-Efficient MSDF Neural Network Inference
Oct 8, 2026Yousef Sadegheih, Dorit Merhof, Muhammad UsmanEfficient Neural Network Inference Early-Exit Neural Networks
Faculty of Informatics and Data Science, University of Regensburg, 93053 Regensburg, Germany
Does per-frame early exit pay? A compute-matched study of dynamic depth for on-device speech enhancement
Sep 24, 2026Clément Laroche, Riccardo MicciniEfficient Neural Network Inference Speech Enhancement
GN A/S, Denmark · Technical University of Denmark, Denmark
Proportional-Fair Resource Allocation and Dual-Threshold Early-Exit Inference for Secure Cooperative Multi-Layer Edge Intelligence
Sep 14, 2026Thai T. Vu, John Le, Tu N. Nguyen +3Edge Inference Fair Resource Allocation
School of Computing and Information Technology, University of Wollongong, Wollongong, NSW, Australia · Department of Computer Science, Kennesaw State University, Georgia, USA · CodeZX Software Company Limited, Vietnam
Adaptive AI: Energy Efficient Multi-exit TinyML on Intelligent Vision Systems at the Edge
Sep 14, 2026Luca Crupi, Lorenzo Lamberti, Alessandro Giusti +1Efficient Neural Network Inference Energy-Efficient ML
IDSIA, SUPSI, Lugano, Switzerland · IIS, ETH Zürich, Switzerland
Theoretical Guarantees for One-Shot Magnitude Pruning and Compute-Adaptive Early Exit
Sep 14, 2026Erdem KoyuncuNeural Network Generalization Early-Exit Neural Networks
Department of Electrical and Computer Engineering, University of Illinois Chicago, IL, USA.
</think> Doesn't Stop Reasoning: Analysis of Spurious CoT Termination
Sep 3, 2026Seunghee Koh, Sungjae Choi, Minchan Kwon +2CoT Reasoning Efficient Language Model Reasoning
Korea Advanced Institute of Science and Technology, South Korea
Adaptive-WAM: Quality-Guided Early-Exit Planning from Intermediate Video-Diffusion Features
Aug 6, 2026Sining Ang, Yuguang Yang, Yan WangEfficient Inference Diffusion Transformer
Institute for AI Industry Research (AIR), Tsinghua University · Department of Automation, University of Science and Technology of China · School of Electronic Information Engineering, Beihang University
Diversity is Not Ambiguity: Toward Accurate and Efficient Ambiguity Detection for Open-Domain QA
Aug 4, 2026Jiwon Lee, Yong-chan Park, Jungin Hong +1Question Answering Open-Domain QA
Seoul National University
Stop When Memory Suffices: Evidence-Conditioned Progressive Execution for LLM Agents
Aug 2, 2026Yidan Lin, Kaixiang Wang, Jiong Lou +1LLM Inference Efficiency Efficient Inference
Real-Time Visual Obstruction Detection in Surgical Augmented Reality
Jul 31, 2026Shih-Chin Yang, Yanming Xiu, Hanting Ye +3Efficient Inference Medical VLMs
Department of Computer Science, Duke University · Department of Electrical and Computer Engineering, Duke University
BLADE: Boundary-Expanded and Layer-Adaptive Dynamic Exit for Efficient LLM Reasoning
Jul 31, 2026Keshu Fu, Keqin Peng, Jun Bai +6Efficient Language Model Reasoning Efficient Language Model Inference
School of Computer Science and Engineering, Beihang University · Beijing Institute for General Artificial Intelligence, BIGAI · Academy for Advanced Interdisciplinary Studies, Peking University +1
Fast Wireless Foundation Models with Early-Exits
Jun 28, 2026Omar Mashaal, Hatem Abou-ZeidWireless Foundation Models Efficient Inference
Department of Electrical and Software Engineering, University of Calgary, Calgary, AB T2N 1N4, Canada
Early-Exit Graph Neural Networks for Link Prediction
Jun 20, 2026Roman Knyazhitskiy, Andrea Giuseppe Di FrancescoEfficient Neural Network Inference Graph Neural Networks
University of Cambridge · Sapienza University of Rome
Think Less, Act Early: Reinforced Latent Reasoning with Early Exit in Vision-Language-Action Models
Jun 13, 2026Dianqiao Lei, Lianlei ShanEfficient Inference Efficient VLA Model Inference
1Tsinghua University, Beijing, China.
ViT-FREE: Efficient Face Recognition via Early Exiting and Synthetic Adaptation
Jun 10, 2026Tahar Chettaoui, Guray Ozgur, Eduarda Caldeira +2Inference-Time Optimization Efficient ViTs
Fraunhofer Institute for Computer Graphics Research IGD, Germany · Department of Computer Science, TU Darmstadt, Germany
TabSwift: An Efficient Tabular Foundation Model with Row-Wise Attention
Jun 5, 2026Si-Yang Liu, Han-Jia YeTabular Foundation Models Efficient Inference
School of Artificial Intelligence, Nanjing University, China · National Key Laboratory for Novel Software Technology, Nanjing University, China.
t-gems: text-guided exit modules for decreasing clip image encoder
May 17, 2026Alberto Presta, Grzegorz Stefanski, Michal Byra +1Cross-Modal Learning Vision-Language Models
Samsung AI Center Warsaw · Institute of Fundamental Technological Research, PAS, Warsaw
Trajectory-Aware Adaptive Inference in Object Detection Models
May 12, 2026Grigorios Papanikolaou, Ioannis Kontopoulos, Giannis Spiliopoulos +2Adaptive Inference Dynamic Neural Networks
Department of Electrical and Computer Engineering, National Technical University of Athens, Greece · Department of Product and Systems Design Engineering, University of the Aegean, Syros, Greece
Amortized-Precision Quantization for Early-Exit Vision Transformers
May 8, 2026Rui Fang, Hsi-Wen Chen, Ming-Syan ChenMixed-Precision Quantization Dynamic Neural Networks
National Taiwan University
Balancing Stability and Plasticity in Sequentially Trained Early-Exiting Neural Networks
May 6, 2026Alaa Zniber, Ouassim Karrakchou, Mounir GhoghoAdaptive Inference Dynamic Neural Networks
⋆TICLab, International University of Rabat, Morocco · ‡ School of Electronic and Electrical Engineering, University of Leeds, UK
LEAP: Layer-wise Exit-Aware Pretraining for Efficient Transformer Inference
May 1, 2026Shashank Kapadia, Deep Naryan Mishra, Sujal Reddy Alugubelli +4Efficient Transformer Inference Dynamic Neural Networks
Walmart Inc. · Sam’s Club
A Comparative Analysis on the Performance of Upper Confidence Bound Algorithms in Adaptive Deep Neural Networks
Apr 27, 2026Grigorios Papanikolaou, Ioannis Kontopoulos, Konstantinos TserpesMulti-Armed Bandits Dynamic Neural Networks
Department of Electrical and Computer Engineering National Technical University of Athens, Greece
Progressive Approximation in Deep Residual Networks: Theory and Validation
Apr 27, 2026Wei Wang, Xiao-Yong Wei, Qing LiEfficient Neural Network Inference Residual Learning
Department of Computing, the Hong Kong Polytechnic University.
EX-FIQA: Leveraging Intermediate Early eXit Representations from Vision Transformers for Face Image Quality Assessment
Apr 21, 2026Guray Ozgur, Tahar Chettaoui, Eduarda Caldeira +4Efficient Transformer Inference Efficient Inference
Fraunhofer Institute for Computer Graphics Research IGD, Darmstadt, Germany · Department of Computer Science, TU Darmstadt, Germany
River-LLM: Large Language Model Seamless Exit Based on KV Share
Apr 20, 2026Yingtao Shen, An ZouDecoder-Only Language Models KV-Cache Management
Shanghai Jiao Tong University Shanghai, China
Step-GRPO: Internalizing Dynamic Early Exit for Efficient Reasoning
Apr 18, 2026Benteng Chen, Weida Wang, Shufei Zhang +2RL for Language Model Reasoning LLM Inference Acceleration
East China Normal University · Shanghai AI Laboratory · Fudan University +1
When Do Early-Exit Networks Generalize? A PAC-Bayesian Theory of Adaptive Depth
Apr 17, 2026Dongxin Guo, Jikun Wu, Siu Ming YiuNeural Network Generalization PAC-Bayesian Generalization Bounds
The University of Hong Kong, Hong Kong, China · Brain Investing Limited, Hong Kong, China
A Comparative Study of CNN Optimization Methods for Edge AI: Exploring the Role of Early Exits
Apr 16, 2026Nekane Fernandez, Ivan Valdes, Steven Van Vaerenbergh +2Efficient Inference Edge Inference
Nekane Fernandez, Ivan Valdes, Idoia de la Iglesia and Julen Arratibel are with the Dept. of Connected and Distributed Intelligence, Ikerlan Technology Research Centre, Arrasate, Spain · Nekane Fernandez and Steven Van Vaerenbergh are with the Dept. of Mathematics, Statistics and Computation, Universidad de Cantabria, Santander, Spain
Channel-Adaptive Edge AI: Maximizing Inference Throughput by Adapting Computational Complexity to Channel States
Mar 3, 2026Jierui Zhang, Jianhao Huang, Kaibin HuangAdaptive Inference Edge Inference
Department of Electrical and Computer Engineering, The University of Hong Kong, Hong Kong
Hardware-Algorithm Co-Optimization of Early-Exit Neural Networks for Multi-Core Edge Accelerators
Dec 4, 2025Alaa Zniber, Arne Symons, Ouassim Karrakchou +2Efficient Neural Network Inference Dynamic Neural Networks
TICLab, International University of Rabat, Morocco · MICAS, KU Leuven, Belgium

Oct 8, 2026, cs.ARJ/K move · Enter open · S save

DEX: Digit-Level Early Exit for Energy-Efficient MSDF Neural Network Inference

Yousef Sadegheih, Dorit Merhof, Muhammad Usman

Faculty of Informatics and Data Science, University of Regensburg, 93053 Regensburg, Germany

Efficient Neural Network Inference Early-Exit Neural Networks U-Net

Open paper AskPDF arXiv

U-Net inference for brain-tumor segmentation requires billions of multiply-accumulate operations, motivating hardware that can reduce computation dynamically rather than relying only on fixed precision or static model compression. Most-significant-digit-first (MSDF) arithmetic exposes the leading digits of a result during computation, enabling output-dependent decisions before the full value is generated. This paper presents an MSDF accelerator for quantized U-Net segmentation with a two-stage grouped processing element supporting signed INT8 operands and in-stream bias accumulation. Four runtime mechanisms operate directly on the output digit stream: exact early negative detection (END) in ReLU layers, exact sign-only decision making in the segmentation head, calibrated low-order-digit skipping, and calibrated pruning. The two approximate mechanisms are selected offline under an accuracy constraint, while execution requires only lightweight control and does not modify the stored weights. On a residual U-Net trained with nnU-Net for BraTS, the proposed mechanisms reduce digit cycles by 38.38% while achieving a mean Dice score of 80.58% on 73 held-out cases, compared with 81.20% for the floating-point model; the exact mechanisms alone reduce cycles by 18.79% without altering the quantized output. Synthesized in 45~~nm, the processing element operates at 500~~MHz, occupies 0.858~~mm $^2$ , and consumes 0.726~~mJ per

192\times192

patch under switching-activity-annotated power analysis. A projected eight-output accelerator with shared activation delivery achieves 16.6~~ms latency and 1.67~~mJ per patch.

Momentum

Related topics

Latest papers 31