Skip to search formSkip to main contentSkip to account menu

A land-based tank aquaculture system coupled with effluent treatment ponds: Effect of stocking density on the growth performance of gibel carp (Carassius gibelio)

Publisher (opens in a new tab)

M3Rec: A Context-Aware Offline Meta-Level Model-Based Reinforcement Learning Approach for Cold-Start Recommendation

This article addresses the cold-start challenge in the RL-based recommender systems by proposing a novel context-aware offline meta-level model-based RL approach for user adaptation, and introduces a mutual information constraint between the user model and recommendation agent.
ACM (opens in a new tab)

Double Wins: Boosting Accuracy and Efficiency of Graph Neural Networks by Reliable Knowledge Distillation

A novel Reliable Knowledge Distillation framework for MLP optimization (RKDMLP) is proposed, which shows strong promise in achieving a “sweet point” in co-optimizing model accuracy and efficiency.
IEEE (opens in a new tab)

Chain-of-Query: Unleashing the Power of LLMs in SQL-Aided Table Understanding via Multi-Agent Collaboration

CoQ adopts natural-language-style representations of table schemas to abstract away structural noise and enhance understanding and achieves substantial accuracy improvements and significantly lowers invalid SQL rates compared to prior generic LLM-based, SQL-aided, and hybrid baselines, confirming its superior effectiveness in table understanding.
arXiv (opens in a new tab)

Addressing Delayed Feedback in Conversion Rate Prediction: A Domain Adaptation Approach

This work proposes a simple framework that redefines CVR prediction under delayed feedback as an unsupervised domain adaptation (UDA) problem, by integrating existing click-through rate (CTR) or CVR models with UDA algorithms.
IEEE (opens in a new tab)

Exploration into RL-based Language Model Finetuning

The report dives into various methods of training LLM’s on preference finetuning under a reinforcement learning point of view, and starts with a baseline approach running SFT on smoltalk + DPO on ultrafeedback, then dives into implementing a novel approach AGRO which claims to set up a uniformed framework for most existing methods on RL finetuning.

ERNIE 5.0 Technical Report

ERNIE 5.0 represents the first production-scale realization of a trillion-parameter unified autoregressive model that supports both multimodal understanding and generation and systematically addresses the challenges of scaling reinforcement learning to unified foundation models.
arXiv (opens in a new tab)

Fairness-Aware Mutual Information for Multimodal Recommendation

This work proposes a modality-guided representation learning framework using fairness-aware mutual information to disentangle sensitive and non-sensitive information from modal embeddings, and adopts a dual mutual information objective to decompose modal embeddings.
IEEE (opens in a new tab)
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy (opens in a new tab), Terms of Service (opens in a new tab), and Dataset License (opens in a new tab)