Multimodal Large Language Models (MLLMs) in healthcare suffer from severe confirmation bias, often hallucinating visual details to support initial, potentially erroneous diagnostic hypotheses. Existing Chain-of-Thought (CoT) approaches lack intrinsic correction mechanisms, rendering them vulnerable to error propagation. To bridge this gap, we propose Dialectic-Med, a multi-agent framework that enforces diagnostic rigor through adversarial dialectics. Unlike static consensus models, Dialectic-Med orchestrates a dynamic interplay between three role-specialized agents: a proponent that formulates diagnostic hypotheses; an opponent equipped with a novel visual falsification module that actively retrieves contradictory visual evidence to challenge the Proponent; and a mediator that resolves conflicts via a weighted consensus graph. By explicitly modeling the cognitive process of falsification, our framework guarantees that diagnostic reasoning is tightly grounded in verified visual regions. Empirical evaluations on MIMIC-CXR-VQA, VQA-RAD, and PathVQA demonstrate that Dialectic-Med not only achieves state-of-the-art performance but also fundamentally enhances the trustworthiness of the reasoning process. Beyond accuracy, our approach significantly enhances explanation faithfulness and decisively mitigates hallucinations, establishing a new standard over single-agent baselines.

Language agents are increasingly used as social simulators, but most are evaluated by local utterance plausibility rather than their ability to reproduce longitudinal discourse dynamics. We introduce PRISM, a calibrated language-agent framework for forecasting and counterfactual simulation of social media threads. PRISM separates three components that are often conflated in prompt-only agents: latent affective evolution, participation control, and language generation. It models affect with a profile-conditioned stochastic differential equation, selects participation through a partially observed activation policy, and uses a language model as a controlled generator conditioned on the resulting state. Agent profiles are instantiated with MBTI-like anchors, but treated only as behavioral priors rather than psychological ground truth. On 47,207 post-rooted discussion threads spanning 159 trending topics and 356,593 hierarchical comments, PRISM outperforms homogeneous, random, continuous-trait, prompt-only, and memory-only baselines in trajectory alignment, held-out thread forecasting, and blind human evaluation. Reliability analysis further validates LLM-derived weak labels against human and independent-model annotations. These results suggest that calibrated state-space agents offer a more reliable foundation for online discourse simulation than persona prompting alone.

Large Language Models (LLMs) have achieved considerable success in text-based mathematical reasoning, yet their potential remains underexplored in the multimodal mathematics domain where joint text and image understanding is imperative. A key bottleneck hindering progress is the scarcity of high-quality, genuinely multimodal benchmarks. To address this gap, we construct a unified benchmark by consolidating and curating three public multimodal mathematics datasets. We subsequently propose the UniMath-CoT framework, which establishes a robust performance baseline by combining Chain-of-Thought (CoT) principles with efficient Supervised Fine-Tuning (SFT) based on Low-Rank Adaptation (LoRA). Furthermore, to bolster the model’s reasoning robustness, we introduce an innovative verification mechanism, AARI (Answer Affirmation by Re-Inference), which leverages a specialized re-inference protocol to have the model self-scrutinize and validate its initial conclusions. Our comprehensive experiments show that this integrated strategy substantially boosts performance, surpassing a wide range of open-source models and markedly closing the gap with leading proprietary systems.