The growing number of generative AI-based dialogue systems has made their evaluation a crucial challenge. This paper presents our contribution to this important problem through the Dialogue System Technology Challenge (DSTC-12, Track 1), where we developed models to predict dialogue-level, dimension-specific scores. Given the constraint of using relatively small models (i.e. fewer than 13 billion parameters) our work follows two main strategies: employing Language Models (LMs) as evaluators through prompting, and training encoder-based classification and regression models. Our results show that while LM prompting achieves only modest correlations with human judgments, it still ranks second on the test set, outperformed only by the baseline. The regression and classification models, with significantly fewer parameters, demonstrate high correlation for some dimensions on the validation set. Although their performance decreases on the test set, it is important to note that the test set contains annotations with significantly different score ranges for some of the dimensions with respect to the train and validation sets.

Large language models (LLMs) are increasingly evident for accurate question answering across various domains. However, rigorous evaluation of their performance on complex question-answering (QA) capabilities is essential before deployment in real-world biomedical and healthcare applications. This paper presents our approach to the MedHopQA track of the BioCreative IX shared task, which focuses on multi-hop biomedical question answering involving diseases, genes, and chemicals. We adopt a supervised fine-tuning strategy leveraging LLaMA 3 8B, enhanced with a curated biomedical question-answer dataset compiled from external sources including BioASQ, MedQuAD, and TREC. Three experimental setups are explored: fine-tuning on combined short and long answers, short answers only, and long answers only. While our models demonstrate strong domain understanding, achieving concept-level accuracy scores of up to 0.8, their Exact Match (EM) scores remain significantly lower, particularly in the test phase. We introduce a two-stage inference pipeline for precise short-answer extraction to mitigate verbosity and improve alignment with evaluation metrics. Despite partial improvements, challenges persist in generating strictly formatted outputs. Our findings highlight the gap between semantic understanding and exact answer evaluation in biomedical LLM applications, motivating further research in output control and post-processing strategies.

A criminal judicial opinion represents the judge's disposition of a case, including the decision rationale and sentencing. Automatically generating such opinions can assist in analyzing sentencing consistency and provide judges with references to similar past cases. However, current research typically approaches this task by dividing it into two isolated subtasks: legal reasoning and sentencing prediction. This separation often leads to inconsistency between the reasoning and predictions, failing to meet real-world judicial requirements. Furthermore, prior studies rely on manually curated knowledge to enhance applicability, yet such methods remain limited in practical deployment. To address these limitations and better align with legal practice, we propose a new LegalAI task: Judicial Opinion Generation, which simultaneously produces both legal reasoning and sentencing decisions. To achieve this, we introduce LegalChainReasoner, a framework that applies structured legal chains to guide the model through comprehensive case assessments. By integrating factual premises, composite legal conditions, and sentencing conclusions, our approach ensures flexible knowledge injection and end-to-end opinion generation. Experiments on two real-world and open-source Chinese legal case datasets demonstrate that our method outperforms baseline models.

Large Language Models (LLMs) have shown strong capabilities in solving problems across domains, including graph-related tasks traditionally addressed by symbolic or algorithmic methods. In this work, we present a framework for structured context injection, where task-specific information is systematically embedded in the input to guide LLMs in solving a wide range of graph problems. Our method does not require fine-tuning of LLMs, making it cost-efficient and lightweight. We observe that certain graph reasoning tasks remain challenging for LLMs unless they are mapped to conceptually grounded representations. However, achieving such mappings through fine-tuning or repeated multi-step querying can be expensive and inefficient. Our approach offers a practical alternative by injecting structured context directly into the input, enabling the LLM to implicitly align the task with grounded conceptual spaces. We evaluate the approach on multiple graph tasks using both lightweight and large models, highlighting the trade-offs between accuracy and computational cost. The results demonstrate consistent performance improvements, showing that structured input context can rival or surpass more complex approaches. Our findings underscore the value of structured context injection as an effective and scalable strategy for graph understanding with LLMs.

The rapid growth of large language models (LLMs) has heightened the demand for accurate detection of AI-generated text, particularly in languages like Chinese, where subtle linguistic nuances pose significant challenges to current methods. In this study, we conduct a systematic comparison of encoder-based Transformers (Chinese BERT-large and RoBERTa-wwm-ext-large), a decoder-only LLM (Alibaba's Qwen2.5-7B/DeepSeek-R1-Distill-Qwen-7B fine-tuned via Low-Rank Adaptation, LoRA), and a FastText baseline using the publicly available dataset from the NLPCC 2025 Chinese AI-Generated Text Detection Task. Encoder models were fine-tuned using a novel prompt-based masked language modeling approach, while Qwen2.5-7B was adapted for classification with an instruction-format input and a lightweight classification head trained via LoRA. Experiments reveal that although encoder models nearly memorize training data, they suffer significant performance degradation under distribution shifts (RoBERTa: 76.3% test accuracy; BERT: 79.3%). FastText demonstrates surprising lexical robustness (83.5% accuracy) yet lacks deeper semantic understanding. In contrast, the LoRA-adapted Qwen2.5-7B achieves 95.94% test accuracy with balanced precision-recall metrics, indicating superior generalization and resilience to dataset-specific artifacts. These findings underscore the efficacy of decoder-based LLMs with parameter-efficient fine-tuning for robust Chinese AI-generated text detection. Future work will explore next-generation Qwen3 models, distilled variants, and ensemble strategies to enhance cross-domain robustness further.

Recently, Omni-modal large language models (OLLMs) have sparked a new wave of research, achieving impressive results in tasks such as audio-video understanding and real-time environment perception. However, hallucination issues still persist. Similar to the bimodal setting, the priors from the text modality tend to dominate, leading OLLMs to rely more heavily on textual cues while neglecting visual and audio information. In addition, fully multimodal scenarios introduce new challenges. Most existing models align visual or auditory modalities with text independently during training, while ignoring the intrinsic correlations between video and its corresponding audio. This oversight results in hallucinations when reasoning requires interpreting hidden audio cues embedded in video content. To address these challenges, we propose OmniDPO, a preference-alignment framework designed to mitigate hallucinations in OLLMs. Specifically, OmniDPO incorporates two strategies: (1) constructing text-preference sample pairs to enhance the model's understanding of audio-video interactions; and (2) constructing multimodal-preference sample pairs to strengthen the model's attention to visual and auditory information. By tackling both challenges, OmniDPO effectively improves multimodal grounding and reduces hallucination. Experiments conducted on two OLLMs demonstrate that OmniDPO not only effectively mitigates multimodal hallucinations but also significantly enhances the models' reasoning capabilities across modalities. All code and datasets will be released upon paper acceptance.

Higher education faces growing challenges in delivering personalized, scalable, and pedagogically coherent learning experiences. This study introduces a structured framework for designing an AI-powered Learning Management System (AI-LMS) that integrates generative and conversational AI to support adaptive, interactive, and learner-centered instruction. Using a design-based research (DBR) methodology, the framework unfolds through five phases: literature review, SWOT analysis, development of ethical-pedagogical principles, system design, and instructional strategy formulation. The resulting AI-LMS features modular components -- including configurable prompts, adaptive feedback loops, and multi-agent conversation flows -- aligned with pedagogical paradigms such as behaviorist, constructivist, and connectivist learning theories. By combining AI capabilities with human-centered design and ethical safeguards, this study advances a practical model for AI integration in education. Future research will validate and refine the system through real-world implementation.

Large language models (LLMs) have shown strong potential in recommendation tasks due to their strengths in language understanding, reasoning and knowledge integration. These capabilities are especially beneficial for review-based recommendation, which relies on semantically rich user-generated texts to reveal fine-grained user preferences and item attributes. However, effectively incorporating reviews into LLM-based recommendation remains challenging due to (1) inefficient to dynamically utilize user reviews under LLMs' constrained context windows, and (2) lacking effective mechanisms to prioritize reviews most relevant to the user's current decision context. To address these challenges, we propose RevBrowse, a review-driven recommendation framework inspired by the "browse-then-decide" decision process commonly observed in online user behavior. RevBrowse integrates user reviews into the LLM-based reranking process to enhance its ability to distinguish between candidate items. To improve the relevance and efficiency of review usage, we introduce PrefRAG, a retrieval-augmented module that disentangles user and item representations into structured forms and adaptively retrieves preference-relevant content conditioned on the target item. Extensive experiments on four Amazon review datasets demonstrate that RevBrowse achieves consistent and significant improvements over strong baselines, highlighting its generalizability and effectiveness in modeling dynamic user preferences. Furthermore, since the retrieval-augmented process is transparent, RevBrowse offers a certain level of interpretability by making visible which reviews influence the final recommendation.

Efficiently steering generative models toward pharmacologically relevant regions of chemical space remains a major obstacle in molecular drug discovery under low-data regimes. We present VECTOR+: Valid-property-Enhanced Contrastive Learning for Targeted Optimization and Resampling, a framework that couples property-guided representation learning with controllable molecule generation. VECTOR+ applies to both regression and classification tasks and enables interpretable, data-efficient exploration of functional chemical space. We evaluate on two datasets: a curated PD-L1 inhibitor set (296 compounds with experimental $IC_{50}$ values) and a receptor kinase inhibitor set (2,056 molecules by binding mode). Despite limited training data, VECTOR+ generates novel, synthetically tractable candidates. Against PD-L1 (PDB 5J89), 100 of 8,374 generated molecules surpass a docking threshold of $-15.0$ kcal/mol, with the best scoring $-17.6$ kcal/mol compared to the top reference inhibitor ($-15.4$ kcal/mol). The best-performing molecules retain the conserved biphenyl pharmacophore while introducing novel motifs. Molecular dynamics (250 ns) confirm binding stability (ligand RMSD < $2.5$ angstroms). VECTOR+ generalizes to kinase inhibitors, producing compounds with stronger docking scores than established drugs such as brigatinib and sorafenib. Benchmarking against JT-VAE and MolGPT across docking, novelty, uniqueness, and Tanimoto similarity highlights the superior performance of our method. These results position our work as a robust, extensible approach for property-conditioned molecular design in low-data settings, bridging contrastive learning and generative modeling for reproducible, AI-accelerated discovery.

Small language models (SLMs) have the capability for text generation and may potentially be used to generate falsified texts online. This study evaluates 14 SLMs (1.7B-14B parameters) including LLaMA, Gemma, Phi, SmolLM, Mistral, and Granite families in generating perceived low and high quality fake news headlines when explicitly prompted, and whether they appear to be similar to real-world news headlines. Using controlled prompt engineering, 24,000 headlines were generated across low-quality and high-quality deceptive categories. Existing machine learning and deep learning-based news headline quality detectors were then applied against these SLM-generated fake news headlines. SLMs demonstrated high compliance rates with minimal ethical resistance, though there were some occasional exceptions. Headline quality detection using established DistilBERT and bagging classifier models showed that quality misclassification was common, with detection accuracies only ranging from 35.2% to 63.5%. These findings suggest the following: tested SLMs generally are compliant in generating falsified headlines, although there are slight variations in ethical restraints, and the generated headlines did not closely resemble existing primarily human-written content on the web, given the low quality classification accuracy.