Huimin Ren

Minghui Zheng, Hongxu Chen, Huimin Ren et al.

Large language models achieve remarkable performance via extended chain-of-thought (CoT) reasoning, yet this lengthy process incurs substantial inference overhead. Existing CoT compression methods struggle with inflexible manual length budgets, computationally expensive multi-stage training pipelines, and fragile scalability restricted to small models. We propose HMPO (Hybrid Median-length Policy Optimization), a cost-effective, single-stage reinforcement learning framework. HMPO efficiently compresses CoT via three synergistic components: an adaptive median-based budget derived from successful rollouts to eliminate manual tuning, a cosine-decay token reward for smooth length penalization, and a multiplicative reward formulation that substantially mitigates trivial reward hacking by strictly prioritizing answer correctness. Trained exclusively on mathematical data, HMPO generalizes seamlessly across math, code, science, and instruction-following tasks. Extensive experiments scaling from 9B to 122B parameters across dense and Mixture-of-Experts (MoE) architectures demonstrate that HMPO achieves 19%--46% token compression with negligible accuracy degradation, all while drastically reducing training costs compared to existing multi-stage baselines.

Huimin Ren, Yan Liang, Baiqiao Su et al.

The ability of Large Language Models (LLMs) to precisely follow complex and fine-grained lexical instructions is a cornerstone of their utility and controllability. However, evaluating this capability remains a significant challenge. Current methods either rely on subjective and costly human evaluation or on automated LLM-as-a-judge systems, which suffer from inherent biases and unreliability. Existing programmatic benchmarks, while objective, often lack the expressiveness to test intricate, compositional constraints at a granular level. To address these limitations, we introduce LexInstructEval, a new benchmark and evaluation framework for fine-grained lexical instruction following. Our framework is built upon a formal, rule-based grammar that deconstructs complex instructions into a canonical <Procedure, Relation, Value> triplet. This grammar enables the systematic generation of a diverse dataset through a multi-stage, human-in-the-loop pipeline and facilitates objective verification via a transparent, programmatic engine. We release our dataset and open-source evaluation tools to facilitate further research into the controllability and reliability of LLMs.

Jian Cheng, Tian Zhang, Shuang Zhang et al.

In knowledge graph construction, a challenging issue is how to extract complex (e.g., overlapping) entities and relationships from a small amount of unstructured historical data. The traditional pipeline methods are to divide the extraction into two separate subtasks, which misses the potential interaction between the two subtasks and may lead to error propagation. In this work, we propose an effective cascade dual-decoder method to extract overlapping relational triples, which includes a text-specific relation decoder and a relation-corresponded entity decoder. Our approach is straightforward and it includes a text-specific relation decoder and a relation-corresponded entity decoder. The text-specific relation decoder detects relations from a sentence at the text level. That is, it does this according to the semantic information of the whole sentence. For each extracted relation, which is with trainable embedding, the relation-corresponded entity decoder detects the corresponding head and tail entities using a span-based tagging scheme. In this way, the overlapping triple problem can be tackled naturally. We conducted experiments on a real-world open-pit mine dataset and two public datasets to verify the method's generalizability. The experimental results demonstrate the effectiveness and competitiveness of our proposed method and achieve better F1 scores under strict evaluation metrics. Our implementation is available at https://github.com/prastunlp/DualDec.

Shaopeng Chen, Chuyue Xie, Huimin Ren et al.

Traditional Deep Learning Recommendation Models (DLRMs) face increasing bottlenecks in performance and efficiency, often struggling with generalization and long-sequence modeling. Inspired by the scaling success of Large Language Models (LLMs), we propose Generative Ranking for Ads at Baidu (GRAB), an end-to-end generative framework for Click-Through Rate (CTR) prediction. GRAB integrates a novel Causal Action-aware Multi-channel Attention (CamA) mechanism to effectively capture temporal dynamics and specific action signals within user behavior sequences. Full-scale online deployment demonstrates that GRAB significantly outperforms established DLRMs, delivering a 3.05% increase in revenue and a 3.49% rise in CTR. Furthermore, the model demonstrates desirable scaling behavior: its expressive power shows a monotonic and approximately linear improvement as longer interaction sequences are utilized.

Sunzhu Li, Jiale Zhao, Miteto Wei et al.

Reinforcement Learning with Verifiable Rewards (RLVR) has driven substantial progress in reasoning-intensive domains like mathematics. However, optimizing open-ended generation remains challenging due to the lack of ground truth. While rubric-based evaluation offers a structured proxy for verification, existing methods suffer from scalability bottlenecks and coarse criteria, resulting in a supervision ceiling effect. To address this, we propose an automated Coarse-to-Fine Rubric Generation framework. By synergizing principle-guided synthesis, multi-model aggregation, and difficulty evolution, our approach produces comprehensive and highly discriminative criteria capable of capturing the subtle nuances. Based on this framework, we introduce RubricHub, a large-scale ($\sim$110k) and multi-domain dataset. We validate its utility through a two-stage post-training pipeline comprising Rubric-based Rejection Sampling Fine-Tuning (RuFT) and Reinforcement Learning (RuRL). Experimental results demonstrate that RubricHub unlocks significant performance gains: our post-trained Qwen3-14B achieves state-of-the-art (SOTA) results on HealthBench (69.3), surpassing proprietary frontier models such as GPT-5. The code and data will be released soon.

Tianfu He, Guochun Chen, Chuishi Meng et al.

People often refer to a place of interest (POI) by an alias. In e-commerce scenarios, the POI alias problem affects the quality of the delivery address of online orders, bringing substantial challenges to intelligent logistics systems and market decision-making. Labeling the aliases of POIs involves heavy human labor, which is inefficient and expensive. Inspired by the observation that the users' GPS locations are highly related to their delivery address, we propose a ubiquitous alias discovery framework. Firstly, for each POI name in delivery addresses, the location data of its associated users, namely Mobility Profile are extracted. Then, we identify the alias relationship by modeling the similarity of mobility profiles. Comprehensive experiments on the large-scale location data and delivery address data from JD logistics validate the effectiveness.