Picture for Yunxin Liu

Yunxin Liu

Institute for AI Industry Research, Shanghai AI Laboratory, Shanghai, China

LoRA-Switch: Boosting the Efficiency of Dynamic LLM Adapters via System-Algorithm Co-design

May 28, 2024
Viaarxiv icon

A Survey of Resource-efficient LLM and Multimodal Foundation Models

Add code
Jan 16, 2024
Viaarxiv icon

Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security

Add code
Jan 10, 2024
Viaarxiv icon

BiSwift: Bandwidth Orchestrator for Multi-Stream Video Analytics on Edge

Dec 25, 2023
Viaarxiv icon

FedTherapist: Mental Health Monitoring with User-Generated Linguistic Expressions on Smartphones via Federated Learning

Add code
Oct 25, 2023
Viaarxiv icon

Accelerating In-Browser Deep Learning Inference on Diverse Edge Clients through Just-in-Time Kernel Optimizations

Sep 16, 2023
Figure 1 for Accelerating In-Browser Deep Learning Inference on Diverse Edge Clients through Just-in-Time Kernel Optimizations
Figure 2 for Accelerating In-Browser Deep Learning Inference on Diverse Edge Clients through Just-in-Time Kernel Optimizations
Figure 3 for Accelerating In-Browser Deep Learning Inference on Diverse Edge Clients through Just-in-Time Kernel Optimizations
Figure 4 for Accelerating In-Browser Deep Learning Inference on Diverse Edge Clients through Just-in-Time Kernel Optimizations
Viaarxiv icon

Empowering LLM to use Smartphone for Intelligent Task Automation

Add code
Sep 09, 2023
Figure 1 for Empowering LLM to use Smartphone for Intelligent Task Automation
Figure 2 for Empowering LLM to use Smartphone for Intelligent Task Automation
Figure 3 for Empowering LLM to use Smartphone for Intelligent Task Automation
Figure 4 for Empowering LLM to use Smartphone for Intelligent Task Automation
Viaarxiv icon

Serving MoE Models on Resource-constrained Edge Devices via Dynamic Expert Swapping

Aug 29, 2023
Viaarxiv icon

Generative Model for Models: Rapid DNN Customization for Diverse Tasks and Resource Constraints

Aug 29, 2023
Figure 1 for Generative Model for Models: Rapid DNN Customization for Diverse Tasks and Resource Constraints
Figure 2 for Generative Model for Models: Rapid DNN Customization for Diverse Tasks and Resource Constraints
Figure 3 for Generative Model for Models: Rapid DNN Customization for Diverse Tasks and Resource Constraints
Figure 4 for Generative Model for Models: Rapid DNN Customization for Diverse Tasks and Resource Constraints
Viaarxiv icon

PatchBackdoor: Backdoor Attack against Deep Neural Networks without Model Modification

Add code
Aug 22, 2023
Figure 1 for PatchBackdoor: Backdoor Attack against Deep Neural Networks without Model Modification
Figure 2 for PatchBackdoor: Backdoor Attack against Deep Neural Networks without Model Modification
Figure 3 for PatchBackdoor: Backdoor Attack against Deep Neural Networks without Model Modification
Figure 4 for PatchBackdoor: Backdoor Attack against Deep Neural Networks without Model Modification
Viaarxiv icon