paper-with-me

Papers

Knowledge boosting during low-latency inference

2024-07-09 · Vidya Srinivas, Malek Itani, Tuochao Chen, Sefik Emre Eskimez, Takuya Yoshioka, Shyamnath Gollakota

Models for low-latency, streaming applications could benefit from the knowledge capacity of larger models, but edge devices cannot run these models due to resource constraints. A possible solution is to transfer hints during inference from a large model running remotely to a small model running on-device. However, this incurs a communication delay that breaks real-time requirements and does not guarantee that both models will operate on the same data at the same time. We propose knowledge boosting, a novel technique that allows a large model to operate on time-delayed input during inference, while still boosting small model performance. Using a streaming neural network that processes 8 ms chunks, we evaluate different speech separation and enhancement tasks with communication delays of up to six chunks or 48 ms. Our results show larger gains where the performance gap between the small and large models is wide, demonstrating a promising method for large-small model collaboration for low-latency applications. Code, dataset, and audio samples available at https://knowledgeboosting.cs.washington.edu/.

📄 PDF Abstract BibTeX arXiv:2407.11055

Code (1)

vysri/knowledge-boosting 공식 구현 pytorch

Tasks

Speech Separation

Similar Papers 제목 키워드 기반

AdaGMLP: AdaBoosting GNN-to-MLP Knowledge Distillation

2024-05-23 · Weigang Lu, Ziyu Guan, Wei Zhao, Yaming Yang

Graph Neural Networks (GNNs) have revolutionized graph-based machine learning, but their heavy computational demands pose challenges for latency-sensitive edge devices in practical industrial applications. In response, a…

Knowledge Distillation

Agentic Cost-Aware Query Planning with Knowledge Distillation for Big Data Analytics

2026-05-18 · Mahdi Naser-Moghadasi arxiv

Query optimization in big data analytics remains computationally expensive, particularly for resource-constrained environments where traditional optimizers fail to satisfy memory and latency constraints. We present an ag…

Knowledge Distillation

QUART-Online: Latency-Free Large Multimodal Language Model for Quadruped Robot Learning

2024-12-20 · Xinyang Tong, Pengxiang Ding, Yiguo Fan, Donglin Wang 외

This paper addresses the inherent inference latency challenges associated with deploying multimodal large language models (MLLM) in quadruped vision-language-action (QUAR-VLA) tasks. Our investigation reveals that conven…

Language ModelingLanguage ModellingVision-Language-Action

Beyond Prediction: Tail-Aware Scheduling for LLM Inference

2026-06-16 · Yueying Li, Yuanfan Chen, Jiayang Chen, Esha Choukse 외 arxiv

LLM serving exhibits extreme length variability, making size-based scheduling difficult in practice. Recent LLM schedulers approximate SJF/SRPT using predicted decode lengths or ranks and primarily report mean-centric me…

LLMBoost: Make Large Language Models Stronger with Boosting

2025-12-26 · Zehao Chen, Tianxiang Ai, Yifei Li, Gongxun Li 외 arxiv

Ensemble learning of LLMs has emerged as a promising alternative to enhance performance, but existing approaches typically treat models as black boxes, combining the inputs or final outputs while overlooking the rich int…

Arithmetic ReasoningEnsemble Learning