CrystalBox: Future-Based Explanations for Input-Driven Deep RL Systems
Authors: Sagar Patel, Sangeetha Abdu Jyothi, Nina Narodytska
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We present an evaluation of Crystal Box. We aim to answer the following questions: Does Crystal Box produce high-fidelity explanations? Is Crystal Box efficient? Does joint training help? What applications can Crystal Box enable? We evaluate Crystal Box using two input-driven environments: Adaptive Bitrate Streaming (ABR) and Congestion Control (CC). |
| Researcher Affiliation | Collaboration | Sagar Patel1, Sangeetha Abdu Jyothi1, 2, Nina Narodytska2 1University of California, Irvine 2VMware Research |
| Pseudocode | No | The paper uses system diagrams and text to describe its methods but does not include a formal pseudocode or algorithm block. |
| Open Source Code | Yes | 2https://github.com/sagar-pa/crystalbox |
| Open Datasets | Yes | All that is required is access to a simulator with training traces, which is publicly available for most input-driven RL environments (Mao et al. 2019). We experiment with the publicly available ABR controller maguro (Patel et al. 2023) deployed on the Puffer Platform (Yan et al. 2020). |
| Dataset Splits | No | The paper mentions training on traces and evaluating on a "held-out set of traces", which implies a train-test split, but does not explicitly describe a separate validation set or specific split percentages for reproduction. |
| Hardware Specification | No | The paper does not provide specific details about the hardware (e.g., CPU, GPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper mentions DRL controllers like 'maguro' and 'Aurora', but does not provide specific version numbers for these or other software dependencies, such as deep learning frameworks or libraries. |
| Experiment Setup | No | The paper describes the overall training procedure and loss function but does not provide specific hyperparameters such as learning rate, batch size, number of epochs, or optimizer settings for the experimental setup. |