# Smoke and Mirrors in Causal Downstream Tasks

arXiv: https://arxiv.org/abs/2405.17151 · HTML: https://mlsanity.com/paper/2405.17151/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 60 references that are in the map.
- First posted on arXiv 27 May 2024.

## Builds on (most cited first, 27 of 27)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Sigmoid Loss for Language Image Pre-Training](https://mlsanity.com/paper/2303.15343/) (2023)
- [Invariant Risk Minimization](https://mlsanity.com/paper/1907.02893/) (2019)
- [Recognition in Terra Incognita](https://mlsanity.com/paper/1807.04975/) (2018)
- [Fairness in Criminal Justice Risk Assessments: The State of the Art](https://mlsanity.com/paper/1703.09207/) (2017)
- [Self-Supervised Learning with Data Augmentations Provably Isolates Content from Style](https://mlsanity.com/paper/2106.04619/) (2021)
- [Weakly-Supervised Disentanglement Without Compromises](https://mlsanity.com/paper/2002.02886/) (2020)
- [Prediction-Powered Inference](https://mlsanity.com/paper/2301.09633/) (2023)
- [Emergence of Invariance and Disentanglement in Deep Representations](https://mlsanity.com/paper/1706.01350/) (2017)
- [Interventional Causal Representation Learning](https://mlsanity.com/paper/2209.11924/) (2022)
- [Weakly supervised causal representation learning](https://mlsanity.com/paper/2203.16437/) (2022)
- [Learning Linear Causal Representations from Interventions under General Nonlinear Mixing](https://mlsanity.com/paper/2306.02235/) (2023)
- [Linear Causal Disentanglement via Interventions](https://mlsanity.com/paper/2211.16467/) (2022)
- [Identifiability Guarantees for Causal Disentanglement from Soft Interventions](https://mlsanity.com/paper/2307.06250/) (2023)
- [Multi-View Causal Representation Learning with Partial Observability](https://mlsanity.com/paper/2311.04056/) (2023)
- [Weakly Supervised Representation Learning with Sparse Perturbations](https://mlsanity.com/paper/2206.01101/) (2022)
- [Cross-Prediction-Powered Inference](https://mlsanity.com/paper/2309.16598/) (2023)
- [Learning latent causal graphs via mixture oracles](https://mlsanity.com/paper/2106.15563/) (2021)
- [Score-based Causal Representation Learning with Interventions](https://mlsanity.com/paper/2301.08230/) (2023)
- [MammalNet: A Large-scale Video Benchmark for Mammal Recognition and Behavior Understanding](https://mlsanity.com/paper/2306.00576/) (2023)
- [The Caltech Fish Counting Dataset: A Benchmark for Multiple-Object Tracking and Counting](https://mlsanity.com/paper/2207.09295/) (2022)
- [MABe22: A Multi-Species Multi-Task Benchmark for Learned Representations of Behavior](https://mlsanity.com/paper/2207.10553/) (2022)
- [A General Framework for Treatment Effect Estimation in Semi-Supervised and High Dimensional Settings](https://mlsanity.com/paper/2201.00468/) (2022)
- [Volumetric Cloud Field Reconstruction](https://mlsanity.com/paper/2311.17657/) (2023)

## Built on by (most cited first, 12 of 12)

- [Unifying Causal Representation Learning with the Invariance Principle](https://mlsanity.com/paper/2409.02772/) (2024)
- [The Landscape of Causal Discovery Data: Grounding Causal Discovery in Real-World Applications](https://mlsanity.com/paper/2412.01953/) (2024)
- [Prediction-Powered Causal Inferences](https://mlsanity.com/paper/2502.06343/) (2025)
- [What is causal about causal models and representations?](https://mlsanity.com/paper/2501.19335/) (2025)
- [The Third Pillar of Causal Analysis? A Measurement Perspective on Causal Representations](https://mlsanity.com/paper/2505.17708/) (2025)
- [CausalVerse: Benchmarking Causal Representation Learning with Configurable High-Fidelity Simulations](https://mlsanity.com/paper/2510.14049/) (2025)
- [Exploratory Causal Inference in SAEnce](https://mlsanity.com/paper/2510.14073/) (2025)
- [MorphGen: Controllable and Morphologically Plausible Generative Cell-Imaging](https://mlsanity.com/paper/2510.01298/) (2025)
- [CaTs and DAGs: Integrating Directed Acyclic Graphs with Transformers for Causally Constrained Predictions](https://mlsanity.com/paper/2410.14485/) (2024)
- [The Narcissus Hypothesis: Descending to the Rung of Illusion](https://mlsanity.com/paper/2509.17999/) (2025)
- [Counterfactual identifiability beyond global monotonicity: non-monotone triangular structural causal models](https://mlsanity.com/paper/2605.04413/) (2026)
- [From Tokens to Policy: Causal and Interpretable Heterogeneous Treatment Effects Identification](https://mlsanity.com/paper/2606.17010/) (2026)

## Abstract (arXiv)

Machine Learning and AI have the potential to transform data-driven scientific discovery, enabling accurate predictions for several scientific phenomena. As many scientific questions are inherently causal, this paper looks at the causal inference task of treatment effect estimation, where the outcome of interest is recorded in high-dimensional observations in a Randomized Controlled Trial (RCT). Despite being the simplest possible causal setting and a perfect fit for deep learning, we theoretically find that many common choices in the literature may lead to biased estimates. To test the practical impact of these considerations, we recorded ISTAnt, the first real-world benchmark for causal inference downstream tasks on high-dimensional observations as an RCT studying how garden ants (Lasius neglectus) respond to microparticles applied onto their colony members by hygienic grooming. Comparing 6 480 models fine-tuned from state-of-the-art visual backbones, we find that the sampling and modeling choices significantly affect the accuracy of the causal estimate, and that classification accuracy is not a proxy thereof. We further validated the analysis, repeating it on a synthetically generated visual data set controlling the causal model. Our results suggest that future benchmarks should carefully consider real downstream scientific questions, especially causal ones. Further, we highlight guidelines for representation learning methods to help answer causal questions in the sciences.