Picture for Aditya Akella

Aditya Akella

UT Austin

On-the-Fly Adaptive Distillation of Transformer to Dual-State Linear Attention

Add code
Jun 12, 2025
Viaarxiv icon

HALoS: Hierarchical Asynchronous Local SGD over Slow Networks for Geo-Distributed Large Language Model Training

Add code
Jun 05, 2025
Viaarxiv icon

ConfigBot: Adaptive Resource Allocation for Robot Applications in Dynamic Environments

Add code
Jan 17, 2025
Figure 1 for ConfigBot: Adaptive Resource Allocation for Robot Applications in Dynamic Environments
Figure 2 for ConfigBot: Adaptive Resource Allocation for Robot Applications in Dynamic Environments
Figure 3 for ConfigBot: Adaptive Resource Allocation for Robot Applications in Dynamic Environments
Figure 4 for ConfigBot: Adaptive Resource Allocation for Robot Applications in Dynamic Environments
Viaarxiv icon

OMEGA: A Low-Latency GNN Serving System for Large Graphs

Add code
Jan 15, 2025
Figure 1 for OMEGA: A Low-Latency GNN Serving System for Large Graphs
Figure 2 for OMEGA: A Low-Latency GNN Serving System for Large Graphs
Figure 3 for OMEGA: A Low-Latency GNN Serving System for Large Graphs
Figure 4 for OMEGA: A Low-Latency GNN Serving System for Large Graphs
Viaarxiv icon

TrainMover: Efficient ML Training Live Migration with No Memory Overhead

Add code
Dec 17, 2024
Figure 1 for TrainMover: Efficient ML Training Live Migration with No Memory Overhead
Figure 2 for TrainMover: Efficient ML Training Live Migration with No Memory Overhead
Figure 3 for TrainMover: Efficient ML Training Live Migration with No Memory Overhead
Figure 4 for TrainMover: Efficient ML Training Live Migration with No Memory Overhead
Viaarxiv icon

C3: Learning Congestion Controllers with Formal Certificates

Add code
Dec 14, 2024
Figure 1 for C3: Learning Congestion Controllers with Formal Certificates
Figure 2 for C3: Learning Congestion Controllers with Formal Certificates
Figure 3 for C3: Learning Congestion Controllers with Formal Certificates
Figure 4 for C3: Learning Congestion Controllers with Formal Certificates
Viaarxiv icon

Read-ME: Refactorizing LLMs as Router-Decoupled Mixture of Experts with System Co-Design

Add code
Oct 24, 2024
Figure 1 for Read-ME: Refactorizing LLMs as Router-Decoupled Mixture of Experts with System Co-Design
Figure 2 for Read-ME: Refactorizing LLMs as Router-Decoupled Mixture of Experts with System Co-Design
Figure 3 for Read-ME: Refactorizing LLMs as Router-Decoupled Mixture of Experts with System Co-Design
Figure 4 for Read-ME: Refactorizing LLMs as Router-Decoupled Mixture of Experts with System Co-Design
Viaarxiv icon

CONGO: Compressive Online Gradient Optimization with Application to Microservices Management

Add code
Jul 08, 2024
Viaarxiv icon

HawkVision: Low-Latency Modeless Edge AI Serving

Add code
May 29, 2024
Viaarxiv icon

FFN-SkipLLM: A Hidden Gem for Autoregressive Decoding with Adaptive Feed Forward Skipping

Add code
Apr 05, 2024
Viaarxiv icon