Research map / Vision-language and multimodal / Human and embodied modeling
Multimodal and affective interaction modeling
Models multimodal interactions, emotions, and social cues with datasets and information decomposition.
87 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Human and embodied modeling, keyword cluster “computer vision · humans · abstract”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
39 papers in 2023–24 and 27 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 7.3% to 4.0%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Building a Large Scale Dataset for Image Emotion Recognition: The Fine Print and The Benchmark (AAAI 2016 · 322 citations)
- 2017 · Connecting Look and Feel: Associating the Visual and Tactile Properties of Physical Materials (CVPR 2017 · 129 citations)
- 2018 · Thinking in PolAR Pictures: Using Rotation-Friendly Mental Images to Solve Leiter-R Form Completion (AAAI 2018 · 1 citation)
- 2019 · Connecting Touch and Vision via Cross-Modal Prediction (CVPR 2019 · 133 citations)
- 2020 · The Hateful Memes Challenge: Detecting Hate Speech in Multimodal Memes (NeurIPS 2020 · 145 citations)
- 2021 · ArtEmis: Affective Language for Visual Art (CVPR 2021 · 146 citations)
- 2022 · NumHTML: Numeric-Oriented Hierarchical Transformer Model for Multi-Task Financial Forecasting (AAAI 2022 · 41 citations)
- 2023 · Learning Emotion Representations From Verbal and Nonverbal Communication (CVPR 2023 · 34 citations)
- 2024 · EmbodiedScan: A Holistic Multi-Modal 3D Perception Suite Towards Embodied AI (CVPR 2024 · 78 citations)
- 2025 · Each Fake News Is Fake in Its Own Way: An Attribution Multi-Granularity Benchmark for Multimodal Fake News Detection (AAAI 2025 · 21 citations)
- 2026 · Multimodal Function Vectors for Visual Relations (ICML 2026 · 0 citations)
Earliest papers here
- Building a Large Scale Dataset for Image Emotion Recognition: The Fine Print and The Benchmark (AAAI 2016 · 322 citations)
- Connecting Look and Feel: Associating the Visual and Tactile Properties of Physical Materials (CVPR 2017 · 129 citations)
- Towards Better Understanding the Clothing Fashion Styles: A Multimodal Deep Learning Approach (AAAI 2017 · 55 citations)
What has been tried
73 papers here report 748 measured changes (a component added, removed or swapped, compared with a baseline): helped 72%, no clear effect 6%, hurt 18%, mixed 3%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| add oriented-box prediction rather than use axis-aligned boxes | ScanNet 18-class detection · AP25 | 70.17 -> 61.87 | Hurt | EmbodiedScan: A Holistic Multi-Modal 3D Perception Suite Towards Embodied AI (CVPR 2024 · 78 citations) |
| add image manipulation latent variable (zIM) | Fakeddit · Accuracy | 0.923 -> 0.925 | Helped | Unveiling Implicit Deceptive Patterns in Multi-Modal Fake News via Neuro-Symbolic Reasoning (AAAI 2024 · 23 citations) |
| add PSCC-NET manipulation features | AMG · F1 | 0.8166 -> 0.8310 | Helped | Each Fake News Is Fake in Its Own Way: An Attribution Multi-Granularity Benchmark for Multimodal Fake News Detection (AAAI 2025 · 21 citations) |
| add Context-Aware Transformer (CAT) | MTL-NAE · mAP | 0.352 -> 0.383 | Helped | Narrative Action Evaluation with Prompt-Guided Multimodal Interaction (CVPR 2024 · 17 citations) |
| Use frame-level crops at scale 2.5 instead of fixed black-mask or fixed without-mask crops | Social Group Level · F1 score average | 0.1452 | Helped | JRDB-Social: A Multifaceted Robotic Dataset for Understanding of Context and Dynamics of Human Interactions Within Social Groups (CVPR 2024 · 16 citations) |
| add difference and Hadamard-product features | Composite · Kendall's τc | 39.3 -> 57.6 | Helped | Polos: Multimodal Metric Learning from Human Feedback for Image Captioning (CVPR 2024 · 16 citations) |
| Add regression loss for continuous parameters instead of cross-entropy-only prediction | DressCode text-to-garment · Accuracy | 79.0 -> 85.0 | Helped | AIpparel: A Multimodal Foundation Model for Digital Garments (CVPR 2025 · 12 citations) |
| Add visual features | YouTube: Speaking Target Identification, Pronoun Coreference Resolution, Mentioned Player Prediction · accuracy | 65.8 -> 72.7; 60.3 -> 65.9; 54.6 -> 58.8 | Helped | Modeling Multimodal Social Interactions: New Challenges and Baselines with Densely Aligned Representations (CVPR 2024 · 9 citations) |
| add direct tactile-text contrastive loss | TVL tactile-text classification · top-1 accuracy | 20.3 -> 36.3 | Helped | A Touch, Vision, and Language Dataset for Multimodal Alignment (ICML 2024 · 8 citations) |
| Add textual EP02, EP09, or EP10 to a visual EmotionPrompt | combined EmotionPrompt · avg | 0.56 -> 0.46; 0.56 -> 0.45; 0.56 -> 0.45 | Hurt | The Good, The Bad, and Why: Unveiling Emotions in Generative AI (ICML 2024 · 7 citations) |
| Disable test-time optimization | Nymeria · MPJPE (mm) | 84.82 -> 85.93 | Hurt | Ego4o: Egocentric Human Motion Capture and Understanding from Multi-Modal Input (CVPR 2025 · 6 citations) |
| train LLaVA to generate explanatory answers instead of fixed binary labels | LOKI · Acc | 81.6 -> 84.3 | Helped | Spot the Fake: Large Multimodal Model-Based Synthetic Image Detection with Artifact Explanation (NeurIPS 2025 · 6 citations) |
What reviewers will expect
Datasets and baselines named in the 24 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- UR_FUNNY · 2 of 24 papers
- AMASS · 2 of 24 papers
Baselines
- GPT-4o · 2 of 24 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 66 papers; from OpenAlex and the papers' affiliations).
- Shanghai Jiao Tong University · 5 papers
- Tsinghua University · 5 papers
- Google · 5 papers
- Stanford University · 4 papers
- Meta · 4 papers
- The Chinese University of Hong Kong · 3 papers
- Tencent · 3 papers
- University of Illinois, Urbana Champaign · 3 papers
- Georgia Institute of Technology · 3 papers
- University of California, San Diego · 3 papers
Use it
Download BibTeX of 36 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Human-like visual cognition in VLMs (97 papers)
- Human perception modeling (71 papers)
- Abstract visual reasoning benchmarks (63 papers)
- Scientific domain vision datasets (47 papers)
Elsewhere in the topic
- Vision-language pretraining (182 papers)
- Medical vision-language models (130 papers)
- Scene graph generation (126 papers)
- Language-grounded scene text and structure (124 papers)
- fMRI visual decoding (106 papers)
- Multimodal retrieval-augmented modeling (85 papers)
- Language-grounded visual concept learning (83 papers)
- Composed and text-based image retrieval (78 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).