Byzantine Resilient Distributed Multi-Task Learning
Authors: Jiani Li, Waseem Abbas, Xenofon Koutsoukos
NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We conduct three experiments for both regression and classification problems and demonstrate that our approach yields good empirical performance for non-convex models, such as convolutional neural networks. |
| Researcher Affiliation | Academia | Jiani Li, Waseem Abbas, and Xenofon Koutsoukos Department of Electrical Engineering and Computer Science Vanderbilt University, Nashville, TN, USA {jiani.li, waseem.abbas, xenofon.koutsoukos}@vanderbilt.edu |
| Pseudocode | No | The paper describes the steps of the proposed rule in text and mathematical formulas but does not provide pseudocode or a clearly labeled algorithm block. |
| Open Source Code | Yes | Our code is available at https://github.com/Jiani Li/resilient Distributed MTL. |
| Open Datasets | Yes | Human Activity Recognition7: Mobile phone sensor data (accelerometer and gyroscope) is collected from 30 individuals... 7https://archive.ics.uci.edu/ml/datasets/human+activity+recognition+using+ smartphones. Digit Classification: We consider a network of ten agents performing digit classification. Five of the ten agents have access to the MNIST dataset8 [45] (group 1) and the other five have access to the synthetic dataset9 (group 2)... 8http://yann.lecun.com/exdb/mnist. 9https://www.kaggle.com/prasunroy/synthetic-digits. |
| Dataset Splits | No | The paper describes the datasets and their use in experiments but does not explicitly provide training, validation, and test dataset splits with specific percentages or sample counts. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory, or cloud instance types) used for running the experiments. |
| Software Dependencies | No | The paper mentions general techniques like 'mini-batch gradient descent' and 'SGD' but does not specify software dependencies with version numbers (e.g., specific libraries or frameworks like TensorFlow/PyTorch with their versions). |
| Experiment Setup | Yes | At each iteration, Byzantine agents send random values (for each dimension) from the interval [15, 16] for target localization, and [0, 0.1] for the other two case studies. |