# Gaze-Anchored Social Net: Decoding Implicit Relations via Joint Modeling

arXiv: https://arxiv.org/abs/2607.22847 · HTML: https://mlsanity.com/paper/2607.22847/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 20 of its 38 references that are in the map.
- First posted on arXiv 24 Jul 2026.

## Builds on (most cited first, 20 of 20)

- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [FiLM: Visual Reasoning with a General Conditioning Layer](https://mlsanity.com/paper/1709.07871/) (2017)
- [Sharpness-Aware Minimization for Efficiently Improving Generalization](https://mlsanity.com/paper/2010.01412/) (2020)
- [Gradient Surgery for Multi-Task Learning](https://mlsanity.com/paper/2001.06782/) (2020)
- [Gaze360: Physically Unconstrained Gaze Estimation in the Wild](https://mlsanity.com/paper/1910.10088/) (2019)
- [Detecting Attended Visual Targets in Video](https://mlsanity.com/paper/2003.02501/) (2020)
- [Connecting Gaze, Scene, and Attention: Generalized Attention Estimation via Joint Modeling of Gaze and Scene Saliency](https://mlsanity.com/paper/1807.10437/) (2018)
- [End-to-End Human-Gaze-Target Detection with Transformers](https://mlsanity.com/paper/2203.10433/) (2022)
- [Understanding Human Gaze Communication by Spatio-Temporal Graph Reasoning](https://mlsanity.com/paper/1909.02144/) (2019)
- [Gaze-LLE: Gaze Target Estimation via Large-Scale Learned Encoders](https://mlsanity.com/paper/2412.09586/) (2024)
- [ChildPlay: A New Benchmark for Understanding Children's Gaze Behaviour](https://mlsanity.com/paper/2307.01630/) (2023)
- [A Modular Multimodal Architecture for Gaze Target Prediction: Application to Privacy-Sensitive Settings](https://mlsanity.com/paper/2307.05158/) (2023)
- [SALSA: A Novel Dataset for Multimodal Group Behavior Analysis](https://mlsanity.com/paper/1506.06882/) (2015)
- [Object-aware Gaze Target Detection](https://mlsanity.com/paper/2307.09662/) (2023)
- [GOO: A Dataset for Gaze Object Prediction in Retail Environments](https://mlsanity.com/paper/2105.10793/) (2021)
- [F-formation Detection: Individuating Free-standing Conversational Groups in Images](https://mlsanity.com/paper/1409.2702/) (2014)
- [JRDB-Social: A Multifaceted Robotic Dataset for Understanding of Context and Dynamics of Human Interactions Within Social Groups](https://mlsanity.com/paper/2404.04458/) (2024)
- [Attention Flow: End-to-End Joint Attention Estimation](https://mlsanity.com/paper/2001.03960/) (2020)
- [Multi-Modal Gaze Following in Conversational Scenarios](https://mlsanity.com/paper/2311.05669/) (2023)
- [Following Gaze Across Views](https://mlsanity.com/paper/1612.03094/) (2016)

## Abstract (arXiv)

Human gaze does more than point to visual targets; it serves as a subtle indicator of social intent within static images, whereas standard models typically process individuals independently, treating gaze as an i.i.d. quantity or predicting social semantics in isolation. Recent multi-person methods attempt to address this but often treat social relations as rigid, post-hoc classifications decoupled from the gaze estimation process. This oversimplification fails to capture the nuanced nature of social intent, which acts as an underlying driver of gaze behavior rather than a secondary categorical output. We address these limitations by proposing ANCHOR, a target-centric paradigm designed to decode gaze-anchored social intent by modeling the joint distribution of visual attention and latent implicit relations. Our approach surfaces these dependencies as the latent structural scaffolding of gaze behavior. The architecture utilizes a relational attention mechanism to capture fine-grained interpersonal links, leveraging feature-wise modulation for efficient multi-person parsing from a single vision backbone. To stabilize the training of this coupled formulation, we implement an optimization synergy to resolve the inherent conflicts between spatial gaze accuracy and latent social reasoning. This approach ensures robust generalization by seeking stable, flat minima while simultaneously harmonizing competing task gradients. We validate our framework on an extended benchmark featuring dense multi-person annotations and novel social influence rankings. Our results demonstrate state-of-the-art performance and provide the first quantitative evidence that implicit social hierarchies can be robustly disentangled and learned directly from static gaze patterns.