2 citations · 3 across the 14 of their papers we have counts for
7 papers · 1 filter
Learning Chaos Without Seeing Chaos: Extrapolation of Global Dynamics in Autoregressive Transformers
Yilun Liu, Yi Zhang, Ganyu Wu +5
Autoregressive models are trained to predict a system's behavior one step at a time, and recursive generation allows the learned dynamics to unfold over long horizons. To what exte…
A Storage-Retrieval Gap in Parametric Knowledge Graph Memory
Martino M. L. Pulici, Cuong Xuan Chu, Evgeny Kharlamov +1
Graph retrieval-augmented generation places retrieved subgraphs into the model's context window at query time, paying a recurring token cost and exposing source data on every call.…
MADA-RL: Multi-Agent Debate-Aware Reinforcement Learning for Parameter-Efficient Reasoning in Compact Models
Martino M. L. Pulici, Cuong Xuan Chu, Evgeny Kharlamov +3
Large language models achieve strong reasoning performance, but often at prohibitive training cost - a challenge that is especially acute for compact models (…
Memory-R2: Fair Credit Assignment for Long-Horizon Memory-Augmented LLM Agents
Sikuan Yan, Ahmed Bahloul, Ercong Nie +4
Memory-augmented LLM agents enable interactions that extend beyond finite context windows by storing, updating, and reusing information across sessions. However, training such agen…
Routing-Free Mixture-of-Experts
Yilun Liu, Jinru Han, Sikuan Yan +2
Standard Mixture-of-Experts (MoE) models rely on centralized routing mechanisms that introduce rigid inductive biases. We propose Routing-Free MoE which eliminates any hard-coded c…
Parameter-Efficient Routed Fine-Tuning: Mixture-of-Experts Demands Mixture of Adaptation Modules
Yilun Liu, Yunpu Ma, Yuetian Lu +3
Mixture-of-Experts (MoE) benefits from a dynamic routing mechanism among their specialized experts, which existing Parameter- Efficient Fine-Tuning (PEFT) strategies fail to levera…
Why long model-based rollouts are no reason for bad Q-value estimates
Philipp Wissmann, Daniel Hein, Steffen Udluft +1
This paper explores the use of model-based offline reinforcement learning with long model rollouts. While some literature criticizes this approach due to compounding errors, many p…