Large language models can improve recommendation quality by reasoning explicitly over user history and candidate items - for example, extracting a user's preferences or explaining why one item fits better than another - rather than mapping history directly to a ranked list. This reasoning, however, is expensive to repeat on every ranking request and, once produced, is typically consumed once and discarded, leaving it neither reusable across future requests nor easy to inspect or correct as user tastes drift. Our insight is that reasoning does not need to be regenerated at every call if it can instead be compressed once into a compact, structured memory that a lightweight model retrieves from. We propose rEDMRec, which distills a teacher LLM's reasoning into four typed, editable experience channels - long-term preference, short-term context, item-perception, and counterfactual hard-negative comparisons - maintained by an LLM memory controller that performs Add/Delete/Modify/Keep operations and refines entries via K-agent debate. A lightweight student LLM then ranks candidates purely by retrieving from this memory, without invoking the teacher again, decoupling online inference cost from reasoning depth. Across ML-1M, Amazon Beauty, and Steam and ten student backbones, rEDMRec improves HR@1 over zero-shot, few-shot, and RAG on every backbone, and over GraphRAG on most backbones, with Impv up to 13.3% vs. the second-best baseline on ML-1M. Channel ablations show that short-term context is the only channel that helps consistently across capacity tiers, whereas long-term, item-perception, and counterfactual contributions are capacity-dependent (and can reverse on the strongest students); debate-based memory optimization lowers bank duplication by 7.4 percentage points while raising downstream HR@1 by up to +0.029 over six optimization epochs.
Minh Hoang Nguyen, Tung Le, Huy-Tien Nguyen· 0 citations
This work introduces Top-K prompting as a robust training and inference paradigm to better capture diverse, plausible reaction predictions and establishes Top-K, plausibility-aware training as a practical new direction for robust future LLM-based synthesis planning.
B. Zagribelnyy, Ivan D. Ilin, N. Bondarev et al.· 0 citations
This work develops MedUAG, an end-to-end trained unified medical model that achieves strong performance across a wide array of understanding and generation tasks, establishing a competitive baseline and paving the way for next-generation medical multimodal systems.
Zijie Meng, Yuncheng Zhang, Hualiang Wang et al.· 0 citations
The first compute-normalised comparison of five TTS families across five open-ended generation benchmarks spanning medicine, law, finance, general chat, and creative writing is conducted - grounded in a unified framework that decomposes the effectiveness of each method's token budget into exploration and exploitation.
Davide Romano, Kanak Raj, Jerrod Parker et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
A novel, model-feedback-free LRM-DoS paradigm that employs the conflict count derived from an Satisfiability Modulo Theories (SMT) solver as a low-cost external signal to guide the synthesis of inference-heavy Constraint Satisfaction Problem (CSP) instances.
Jian Yang, Zhenqi Feng, Zhaoyang Yu et al.· 0 citations
The proposed Module Level Reward Evolution Framework integrates three mechanisms: reflection-based refinement, hybrid credit assignment, and a merge strategy with rollback, which together improve the effectiveness and robustness of reward optimization.
Chenglin Liu, Xun Wang, Ruishuo Chen et al.· 0 citations
This paper investigates how multilingual medical adaptation reshapes the internal representations of Whisper models through layer-wise encoder analysis, and shows that English medical fine-tuning produces the dominant encoder shift, whereas multilingual continuation largely preserves the adapted representation space.
Souranil Kahali, Rituparna Bose, Abner Hernandez et al.· 0 citations
A neuro-symbolic pipeline is proposed that uses large language models to generate intermediate implicit premises that are translated into logical formulae and used with logical formulae representing explicit premises and explicit claims to show the logical relationships between them (entailment, contradiction, or neutrality).
This paper explores the intersection of AI hallucinations and the question of AI consciousness, examining whether the erroneous outputs generated by large language models (LLMs) could be mistaken for signs of emergent intelligence. AI hallucinations, which are false or unverifiable
statements produced by LLMs, raise significant philosophical and ethical concerns. While these hallucinations may appear as data anomalies, they challenge our ability to discern whether LLMs are merely sophisticated simulators of intelligence or could develop genuine cognitive processes. By
analysing the causes of AI hallucinations, their impact on the perception of AI cognition, and the potential implications for AI consciousness, this paper contributes to the ongoing discourse on the nature of artificial intelligence and its future evolution.
Kristina Šekrst· Journal of Consciousness Stu...· 1 citation
A quantitative account of the failure of majority voting over multiple LLM samples to raise answer accuracy, yet its gain varies erratically: on hard questions it can even backfire.
Lizhuo Zhang, Mengmeng Tang, Chenfeng Long et al.· 0 citations
This paper proposes BFTR (Budget-First Tariff Recommendation), a complete algorithmic framework integrating eight Budget-First strategies, including two original hybrid approaches: Recursive Hybrid (conditional interpolation) and Knapsack-First Hybrid (priority knapsack).
Experiments on MemFuseBench show that MemFuse achieves the best overall performance among the evaluated memory systems under all three LLM settings and consistently improves performance on questions requiring cross-source evidence fusion.
A new method, called CW-Net, translates the reasoning process of an autonomous vehicle’s AI system into understandable concepts that explain its behavior.
MIT News · Artificial Intelligence· news.mit.eduAug 31, 2026
With millions of users across the world, Julia has been used to conduct cutting-edge research and to design new drugs, jet engines, heat pumps, and more.
A new machine-learning framework aims to improve the success rate of computational protein design while moving away from results that reproduce sequences found in nature.
New MIT research could lead to better materials for a fossil-fuel-free process for making the chemical that's essential to fertilizer and other products.
