The paper introduces TRACE-Router, a task-level routing framework for large language models that aligns routing with the unit of supervision using delayed task feedback.

This paper presents AlphaRoute, a multi-objective adaptive search framework for VLSI global routing using Large Language Models as semantic policy optimizers.

This paper introduces the $k$-Step-Central Shortest Path problem to maximize reachability in transportation networks and provides a polynomial-time algorithm for unweighted graphs.

The paper introduces ProbMoE, a probabilistic routing framework that tackles the non-differentiability of top-$k$ routing in Mixture-of-Experts (MoE) models, achieving strong performance with improved…

SwarmHarness introduces a decentralized, incentive-aligned protocol enabling self-organizing compute swarms for AI tasks, eliminating the need for central coordinators or heavy blockchain infrastructu…

The paper introduces RA-MoE, a novel fine-tuning framework that leverages the internal routing structure of Mixture-of-Experts (MoE) models to improve performance on multilingual downstream tasks by a…

The paper introduces GONDOR, a memory-efficient extension of Greedy Best-First Search (GBFS) that enables search continuation under strict memory constraints by periodically compressing the search tre…

This paper introduces Stigmergic Graph Memory (SGM), a method to improve warehouse throughput in many-to-many Multi-Agent Pickup and Delivery (MAPD) by using a bounded, decaying memory layer to record…

MViewRouter proposes a multi-view framework that internalizes geometric equivariance using a Multi-view Alternating Attention mechanism to improve generalization and stabilize training for combinatori…

This paper proposes a new router redesign for Mixture-of-Experts models using Manifold Power Iteration to align router rows with the principal singular directions of associated experts.

The paper introduces R$^2$A, an adversarial attack that uses suffix optimization to mislead black-box LLM routers into consistently selecting expensive, high-capability models.

The paper proposes Predictive Routing Replay (PR2) to stabilize reinforcement learning on Mixture of Experts (MoE) LLMs by predicting and incorporating short-horizon router evolution during training a…

Lodestar is a novel online learning-based request routing system that significantly improves LLM inference efficiency by dynamically assigning incoming requests to the optimal GPU instance to minimize…

The paper proposes Task-Aware Coactivation Grouping (TACG) to significantly reduce communication costs in multi-task MoE inference by grouping experts based on task-specific co-activation patterns, ou…

The paper introduces AsyncTool, a new benchmark designed to evaluate LLM agents' ability to handle multiple, concurrent tasks with delayed tool feedback, demonstrating that asynchronous coordination i…

This paper presents an optimized algorithm for computing cut sets of a path set in graph theory and introduces a vectorized computational framework for property calculations.

The paper proposes RoRo, a rubric-guided process reward framework that improves stepwise model routing by evaluating the quality of intermediate reasoning steps, leading to better performance and cost…