12 papers
LoRASpace: A Pool-Wide Shared Substrate for Static and Dynamic Multi-LoRA Composition
Thamirawaran Sathiyalogeswaran, Navindu De Silva, Chamikara Siriwardane +4
Low-rank adaptation (LoRA) attaches small low-rank matrices to a frozen backbone, and public repositories host many LoRAs per backbone, each for a different concept or task. Combin…
Automated Prefetching for Object Spatial Programming Using Temporal Trace Graphs
Baichuan Li, Jayanaka Dantanarayana, Savini Kashmira +2
Dependent database accesses can serialize application--database round trips on a request's critical path. Planning across traversal steps requires knowing both which relationships…
ContextRender: From Execution Dependencies to Agent Context
Savini Kashmira, Jayanaka L. Dantanarayana, Lingjia Tang +1
LLM agents performing long-horizon tasks accumulate tool results that later steps may need. Passing the full history to every invocation is costly even when it fits within the cont…
SIGIL: Skill Compilation for Reliable and Efficient Agent Execution
Jayanaka Dantanarayana, Savini Kashmira, Lingjia Tang +1
Agent skills describe reusable procedures, but runtime models must still interpret their instructions and coordinate execution. We introduce Skill Compilation, which translates the…
Kernel Forge: An Agent Harness for LLM-based Generation and Optimization of CUDA Kernels
Joshua Brodsky, Dhravid Kumar, Savini Kashmira +4
Machine learning models are increasingly embedded in everyday software, and most of their runtime is spent in a small set of compute kernels such as matrix multiplication, convolut…
Prompt Less, Smile More: MTP with Semantic Engineering in Lieu of Prompt Engineering
Jayanaka L. Dantanarayana, Savini Kashmira, Thakee Nathees +4
AI-Integrated programming is emerging as a foundational paradigm for building intelligent systems with large language models (LLMs). Recent approaches such as Meaning Typed Program…
GraphMend: Code Transformations for Fixing Graph Breaks in PyTorch 2
Savini Kashmira, Jayanaka Dantanarayana, Thamirawaran Sathiyalogeswaran +3
This paper presents GraphMend, a compiler technique that automatically fixes FX graph breaks in PyTorch 2 programs. Although PyTorch 2 introduced TorchDynamo and TorchInductor to e…
GraphRunner: A Multi-Stage Framework for Efficient and Accurate Graph-Based Retrieval
Savini Kashmira, Jayanaka L. Dantanarayana, Krisztián Flautner +2
Conventional Retrieval Augmented Generation (RAG) approaches are common in text-based applications. However, they struggle with structured, interconnected datasets like knowledge g…
SLMEval: Entropy-Based Calibration for Human-Aligned Evaluation of Large Language Models
Roland Daynauth, Christopher Clarke, Krisztian Flautner +2
The LLM-as-a-Judge paradigm offers a scalable, reference-free approach for evaluating language models. Although several calibration techniques have been proposed to better align th…
Ranking Unraveled: Recipes for LLM Rankings in Head-to-Head AI Combat
Roland Daynauth, Christopher Clarke, Krisztian Flautner +2
Deciding which large language model (LLM) to use is a complex challenge. Pairwise ranking has emerged as a new method for evaluating human preferences for LLMs. This approach entai…
Aligning Model Evaluations with Human Preferences: Mitigating Token Count Bias in Language Model Assessments
Roland Daynauth, Jason Mars
The SLAM paper demonstrated that on-device Small Language Models (SLMs) are a viable and cost-effective alternative to API-based Large Language Models (LLMs), such as OpenAI's GPT-…
Guylingo: The Republic of Guyana Creole Corpora
Christopher Clarke, Roland Daynauth, Charlene Wilkinson +2
While major languages often enjoy substantial attention and resources, the linguistic diversity across the globe encompasses a multitude of smaller, indigenous, and regional langua…