Task-Specific Multimodal Question Answering Agents via Confidence Calibration and Incremental Reasoning for QANTA 2026

Intelligent Prompt Construction for Large Language Models in Knowledge-based Visual Question Answering

The Intelligent Prompt Construction Framework (IPCF), which equips an autonomous agent with the ability to dynamically generate task-specific prompts, and achieves performance gains over existing baselines on the OK-VQA and A-OKVQA datasets.

A Hybrid Dual-Branch Retrieval and Chain-of-Thought Reasoning Framework for Multimodal Legal Question Answering

TRACE-QA: Task-routed constraint elimination for auditable multi-agent question answering

The proposed TRACE-QA, a training-free multi-agent protocol that routes each instance to a sparse set of reasoning operators, constructs option-blind necessity constraints, audits every candidate in a structured elimination ledger, revisits risky eliminations through global risk-aware rescue, and aggregates role-specialized judge votes, supports constraint-guided elimination as an auditable alternative to answer-first debate.

L-TQA: A Modular Multi-Agent Architecture for Context-Constrained Tabular Question Answering

Question Answering over Tabular Data requires accurate numerical reasoning and robust schema interpretation, both of which remain challenging for restricted-parameter language models. This paper presents Laura Tabular Question Answer (L-TQA), a multi-agent architecture designed to mitigate context limitations by decoupling semantic interpretation, schema filtering, and executable code generation. The proposed system improves query synthesis through targeted contextual constraints, combining a response-type classifier to enforce output formats with a dynamic column selector to reduce schema noise. Evaluated on the SemEval-2025 Task 8 DataBench benchmark, L-TQA establishes a new state-of-the-art for models with up to 9 billion parameters, achieving 77.01% accuracy on the test split. In addition, the proposed modular design generalizes effectively to larger proprietary models, highlighting its structural robustness and practical value for privacy-preserving, data-driven applications.

A Verifier-Guided Explainable Reasoning Framework with Gold-Anchored QLoRA, Task-Aware Mixture-of-Experts, and Group-Relative RLVR

Large language models (LLMs) show strong reasoning ability, but their explanations can remain inconsistent, weakly grounded, or difficult to verify. We propose a verifier-guided explainable reasoning framework for transparent educational question answering that combines gold-anchored QLoRA, task-aware symbolic routing, and group-relative RLVR. Qwen2.5-3B-Instruct is first adapted with field-weighted QLoRA supervision anchored to authoritative answers. A lightweight router then assigns logic problems to a FOL/Z3 verifier and physics problems to a formula- and unit aware symbolic solver. Verifier feedback is further used to support candidate evaluation, self-revision, and reward construction during RLVR. Candidate responses are evaluated along three complementary dimensions: P1 for answer correctness, P2 for evidence or unit consistency, and P3 for reasoning depth and explainability. At inference, gold-free self-consistency aggregates multiple candidate responses before an optional question-only physics verifier performs conservative system-level correction. On 438 held-out examples, RLVR increases P3 from 50.68% to 72.20%, while hybrid P1 remains approximately stable at 55.94%. Self-consistency improves model only P1 from 48.86% to 50.23%, with symbolic verification providing the remaining hybrid gain. These results indicate that RLVR primarily strengthens explicit reasoning structure, while symbolic verification complements the neural policy by improving answer reliability at the system level.

Generating then Refining for Reliable Knowledge Base Question Answering

Evaluations on standard KBQA benchmarks show that the proposed ARI-KBQA enhances model performance with a reduced search space, especially in complex multi-hop query scenarios.