Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The Shift Toward Open and Reproducible AI Research," Under Review, 2026. doi: 10.48550/arXiv.2606.16974..
Learning Deep Features in Instrumental Variable Regression
Authors: Liyuan Xu, Yutian Chen, Siddarth Srinivasan, Nando de Freitas, Arnaud Doucet, Arthur Gretton
ICLR 2021 | Venue PDF | LLM Run Details | Input Tokens: 29,634 Total number of tokens sent to the LLM as input for this paper's analysis. | Output Tokens: 5,756 Total number of tokens produced by the LLM (including reasoning/thinking tokens) for this paper's analysis.
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate the empirical performance of DFIV in Section 4, covering three settings: a classical demand prediction example from econometrics, a challenging IV setting where the treatment consists of high-dimensional image data, and the problem of oļ¬-policy policy evaluation in reinforcement learning. |
| Researcher Affiliation | Collaboration | Liyuan Xu Gatsby Unit EMAIL Yutian Chen Deep Mind EMAIL Siddarth Srinivasan University of Washington EMAIL Nando de Freitas Deep Mind EMAIL Arnaud Doucet Deep Mind EMAIL Arthur Gretton Gatsby Unit EMAIL |
| Pseudocode | Yes | Algorithm 1 Deep Feature Instrumental Variable Regression |
| Open Source Code | Yes | The code is included in the supplemental material. |
| Open Datasets | Yes | We used the demand design dataset of Hartford et al. (2017) for benchmarking in the low and high-dimensional cases, and we propose a new setting for the high-dimensional case based on the d Sprites dataset (Matthey et al., 2017). In the deep RL context, we also apply DFIV to perform oļ¬-policy policy evaluation (OPE). The algorithms in the ļ¬rst two experiments are implemented using Py Torch (Paszke et al., 2019) and the OPE experiments are implemented using Tensor Flow (Abadi et al., 2015) and the Acme RL framework (Hoļ¬man et al., 2020). The code is included in the supplemental material. |
| Dataset Splits | Yes | We tuned the regularizers Ī»1, Ī»2 as discussed in Appendix A, with the data evenly split for stage 1 and stage 2. |
| Hardware Specification | No | The algorithms in the ļ¬rst two experiments are implemented using Py Torch (Paszke et al., 2019) and the OPE experiments are implemented using Tensor Flow (Abadi et al., 2015) and the Acme RL framework (Hoļ¬man et al., 2020). The code is included in the supplemental material. No specific hardware specifications (e.g., GPU/CPU models, memory) are mentioned for the experiments. |
| Software Dependencies | Yes | The algorithms in the ļ¬rst two experiments are implemented using Py Torch (Paszke et al., 2019) and the OPE experiments are implemented using Tensor Flow (Abadi et al., 2015) and the Acme RL framework (Hoļ¬man et al., 2020). |
| Experiment Setup | Yes | Unless otherwise speciļ¬ed, all neural network-based algorithms are optimized using Adam with learning rate = 0.001, Ī²1 = 0.9, Ī²2 = 0.999 and Īµ = 10 8. For DFIV, we used the structure described in Table 4. The regularizer Ī»1, Ī»2 are both set to 0.1 as a result of the tuning procedure described in Appendix A. |