# DIDER: Discovering Interpretable Dynamically Evolving Relations

arXiv: https://arxiv.org/abs/2208.10592 · HTML: https://mlsanity.com/paper/2208.10592/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 35 references that are in the map.
- First posted on arXiv 22 Aug 2022.

## Builds on (most cited first, 19 of 19)

- [Categorical Reparameterization with Gumbel-Softmax](https://mlsanity.com/paper/1611.01144/) (2016)
- [The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables](https://mlsanity.com/paper/1611.00712/) (2016)
- [Neural Relational Inference for Interacting Systems](https://mlsanity.com/paper/1802.04687/) (2018)
- [The inD Dataset: A Drone Dataset of Naturalistic Road User Trajectories at German Intersections](https://mlsanity.com/paper/1911.07602/) (2019)
- [Sequential Neural Models with Stochastic Layers](https://mlsanity.com/paper/1605.07571/) (2016)
- [Amortized Causal Discovery: Learning to Infer Causal Graphs from Time-Series Data](https://mlsanity.com/paper/2006.10833/) (2020)
- [Learning Disentangled Joint Continuous and Discrete Representations](https://mlsanity.com/paper/1804.00104/) (2018)
- [Learning Robot Skills with Temporal Variational Inference](https://mlsanity.com/paper/2006.16232/) (2020)
- [Temporal Difference Variational Auto-Encoder](https://mlsanity.com/paper/1806.03107/) (2018)
- [Interpretable Social Anchors for Human Trajectory Forecasting in Crowds](https://mlsanity.com/paper/2105.03136/) (2021)
- [SKID RAW: Skill Discovery from Raw Trajectories](https://mlsanity.com/paper/2103.14610/) (2021)
- [Online Multi-Agent Forecasting with Interpretable Collaborative Graph Neural Network](https://mlsanity.com/paper/2107.00894/) (2021)
- [GRIT: Fast, Interpretable, and Verifiable Goal Recognition with Learned Decision Trees for Autonomous Driving](https://mlsanity.com/paper/2103.06113/) (2021)
- [Joint Interaction and Trajectory Prediction for Autonomous Driving using Graph Neural Networks](https://mlsanity.com/paper/1912.07882/) (2019)
- [Domain Knowledge Driven Pseudo Labels for Interpretable Goal-Conditioned Interactive Trajectory Prediction](https://mlsanity.com/paper/2203.15112/) (2022)
- [Extracting Traffic Primitives Directly from Naturalistically Logged Data for Self-Driving Applications](https://mlsanity.com/paper/1709.03553/) (2017)
- [Grounded Relational Inference: Domain Knowledge Driven Explainable Autonomous Driving](https://mlsanity.com/paper/2102.11905/) (2021)
- [Multi-Vehicle Interaction Scenarios Generation with Interpretable Traffic Primitives and Gaussian Process Regression](https://mlsanity.com/paper/1910.03633/) (2019)
- [Dynamic Relational Inference in Multi-Agent Trajectories](https://mlsanity.com/paper/2007.13524/) (2020)

## Built on by (most cited first, 2 of 2)

- [Disentangled Neural Relational Inference for Interpretable Motion Prediction](https://mlsanity.com/paper/2401.03599/) (2024)
- [Online Relational Inference for Evolving Multi-agent Interacting Systems](https://mlsanity.com/paper/2411.01442/) (2024)

## Abstract (arXiv)

Effective understanding of dynamically evolving multiagent interactions is crucial to capturing the underlying behavior of agents in social systems. It is usually challenging to observe these interactions directly, and therefore modeling the latent interactions is essential for realizing the complex behaviors. Recent work on Dynamic Neural Relational Inference (DNRI) captures explicit inter-agent interactions at every step. However, prediction at every step results in noisy interactions and lacks intrinsic interpretability without post-hoc inspection. Moreover, it requires access to ground truth annotations to analyze the predicted interactions, which are hard to obtain. This paper introduces DIDER, Discovering Interpretable Dynamically Evolving Relations, a generic end-to-end interaction modeling framework with intrinsic interpretability. DIDER discovers an interpretable sequence of inter-agent interactions by disentangling the task of latent interaction prediction into sub-interaction prediction and duration estimation. By imposing the consistency of a sub-interaction type over an extended time duration, the proposed framework achieves intrinsic interpretability without requiring any post-hoc inspection. We evaluate DIDER on both synthetic and real-world datasets. The experimental results demonstrate that modeling disentangled and interpretable dynamic relations improves performance on trajectory forecasting tasks.