arXiv:2607. 18973v1 Announce Type: cross Abstract: Textual skills provide a lightweight way to improve frozen language-model agents, but their self-evolution normally requires a stable validation signal.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.
RL-ADA introduces a co‑evolutionary training framework that replaces costly human annotations with world‑feedback rewards derived from interaction outcomes. In this system, a large Customer Support Agent and an Adversarial Customer Agent train together, guided by an automated judge that rewards successful resolution and realistic intent‑concealing utterances, respectively. Applied to a banking support proof of concept, the method eliminates routing errors and doubles the end‑to‑end PASS rate over five cycles, while also revealing a new adversarial strategy called Contextual Camouflage.
arXiv:2605. 28882v2 Announce Type: replace-cross Abstract: With the rapid advancement of large language models, evaluating human-likeness in open-ended conversation has become increasingly important.
PragAlign is a feedback‑guided framework that improves synthetic dialogue generation by iteratively generating, evaluating, and revising conversations to meet specified service context, target intent, and target emotion. Using an LLM‑based evaluator that scores intent alignment, emotion alignment, coherence, fluency, and overall quality, PragAlign achieves a 99.50% acceptance rate on 800 dialogue specifications, outperforming one‑shot and repeated generation without feedback. Human evaluation confirms that intent expression and dialogue flow are reliably recognized, while emotion appropriateness remains more variable.
arXiv:2606. 21097v2 Announce Type: replace-cross Abstract: Deploying highly capable personalized conversational agents in resource-constrained or privacy-sensitive environments remains a significant challenge.
J-Zero introduces a unified Challenger–Solver–Judge self‑evolution framework that operates without any initial data. The Challenger and Solver co‑evolve through adversarial task generation and response improvement, while the Judge adapts using known preference pairs derived from the Solver’s outputs rather than its own scores. Experiments show J‑Zero surpasses baselines by 4.2 points on verifiable tasks and 8.0 points on unverifiable tasks, maintaining improvement over ten iterations versus baseline degradation after two.
J-Zero introduces a unified Challenger–Solver–Judge co‑evolution framework that enables self‑improvement of language models without requiring external supervision. The Challenger generates increasingly difficult tasks, the Solver learns to produce better responses, and the Judge adapts using preference pairs derived from the Solver’s own outputs rather than from its own scores. Experiments show J‑Zero outperforms baselines by an average of 4.2 points on verifiable tasks and 8.0 points on unverifiable tasks, maintaining improvement over ten iterations while baselines degrade after two.