Revisiting SMoE Language Models by Evaluating Inefficiencies with Task Specific Expert Pruningby Soumajyoti Sarkar, Leonard…
Large-scale Urban Facility Location Selection with Knowledge-informed Reinforcement Learningby Hongyuan Su, Yu Zheng, Jingtao Ding,…
CHESS: Optimizing LLM Inference via Channel-Wise Thresholding and Selective Sparsificationby Junhui He, Shangyu Wu, Weidong…
Dataset Distillation from First Principles: Integrating Core Information Extraction and Purposeful Learningby Vyacheslav Kungurtsev, Yuanfang…
Prompt Compression with Context-Aware Sentence Encoding for Fast and Improved LLM Inferenceby Barys Liskavets, Maxim…
Beyond Parameter Count: Implicit Bias in Soft Mixture of Expertsby Youngseog Chung, Dhruv Malik, Jeff…
Compressing VAE-Based Out-of-Distribution Detectors for Embedded Deploymentby Aditya Bansal, Michael Yuhas, Arvind EaswaranFirst submitted to…
Hyper-Compression: Model Compression via Hyperfunctionby Fenglei Fan, Juntong Fan, Dayang Wang, Jingbo Zhang, Zelin Dong,…
TinyAgent: Function Calling at the Edgeby Lutfi Eren Erdogan, Nicholas Lee, Siddharth Jha, Sehoon Kim,…
Lyapunov Neural ODE State-Feedback Control Policiesby Joshua Hang Sai Ip, Georgios Makrygiorgos, Ali MesbahFirst submitted…