# Recognition-Conditioned Reasoning: A Training-Free Multimodal-LLM Pipeline for Fine-Grained Micro-Action Understanding

arXiv: https://arxiv.org/abs/2608.21022 · HTML: https://mlsanity.com/paper/2608.21022/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning.
- Builds on 11 of its 54 references that are in the map.
- First posted on arXiv 21 Aug 2026.

## Placed in

- LLM reasoning / [Spatial and embodied reasoning](https://mlsanity.com/lr/spatial-and-embodied-reasoning/)

## Builds on (most cited first, 11 of 11)

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [MiniCPM-V: A GPT-4V Level MLLM on Your Phone](https://mlsanity.com/paper/2408.01800/) (2024)
- [MLLM-as-a-Judge: Assessing Multimodal LLM-as-a-Judge with Vision-Language Benchmark](https://mlsanity.com/paper/2402.04788/) (2024)
- [Joint Skeletal and Semantic Embedding Loss for Micro-gesture Classification](https://mlsanity.com/paper/2307.10624/) (2023)
- [Micro-gesture Online Recognition using Learnable Query Points](https://mlsanity.com/paper/2407.04490/) (2024)
- [Prototype Learning for Micro-gesture Classification](https://mlsanity.com/paper/2408.03097/) (2024)
- [MM-Gesture: Towards Precise Micro-Gesture Recognition through Multimodal Fusion](https://mlsanity.com/paper/2507.08344/) (2025)
- [Online Micro-gesture Recognition Using Data Augmentation and Spatial-Temporal Attention](https://mlsanity.com/paper/2507.09512/) (2025)
- [Spatial-Temporal Decoupled Adapter for Micro-gesture Online Recognition](https://mlsanity.com/paper/2606.07355/) (2026)
- [Self-supervised Learning Matters: A Simple Ensemble Solution for Micro-Gesture Recognition](https://mlsanity.com/paper/2606.09261/) (2026)
- [Rethinking the Role of Feature Engineering and Learning Strategies in Few-Shot Hidden Emotion Recognition](https://mlsanity.com/paper/2606.31249/) (2026)

## Abstract (arXiv)

Micro-actions are subtle, short, low-amplitude body movements, such as a fidgeting hand or a slight head tilt, that humans perform with little conscious intent yet that reliably leak emotional and psychological state. Understanding them goes beyond assigning a label: a model must also describe which body parts move and reason, faithfully, about why a clip warrants a particular fine-grained category. We present the training-free, prompt-only system that won first place in the fine-grained understanding track (MA-Bench) of the MAC~2026 Micro-Action Challenge, where both fine-tuning and ground-truth supervision are disallowed. Built entirely upon frozen multimodal large language models (MLLMs), the system dynamically routes each of the eight sub-tasks to the MLLM empirically best suited for that task: a discriminative MLLM for closed-ended recognition tasks and a generative MLLM for open-ended description and reasoning tasks. This architecture achieves a statistically significant performance advantage on open-ended tasks, attaining an average score of 2.68 (on a five-point scale) compared to 1.44 for the second-best approach.