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 embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Learning Multiple Tasks with Multilinear Relationship Networks
Authors: Mingsheng Long, ZHANGJIE CAO, Jianmin Wang, Philip S. Yu
NeurIPS 2017 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments show that MRN yields state-of-the-art results on three multi-task learning datasets. |
| Researcher Affiliation | Academia | Mingsheng Long, Zhangjie Cao, Jianmin Wang, Philip S. Yu School of Software, Tsinghua University, Beijing 100084, China EMAIL EMAIL EMAIL |
| Pseudocode | No | The paper describes the algorithm steps in paragraph text within Section 4.2 but does not provide a formally structured pseudocode or algorithm block. |
| Open Source Code | No | Codes and datasets will be released. |
| Open Datasets | Yes | Office-Caltech [12] This dataset is the standard benchmark for multi-task learning and transfer learning. ... Office-Home1 [26] ... Image CLEF-DA2 |
| Dataset Splits | Yes | We conduct model selection for all methods using five-fold cross-validation on the training set. |
| Hardware Specification | No | The paper mentions using Alex Net and VGGnet as base architectures, but does not specify any particular hardware (e.g., GPU models, CPU, memory) used for the experiments. |
| Software Dependencies | No | The paper mentions using Alex Net and VGGnet architectures and De CAF7 features, but does not specify any software versions (e.g., Python, TensorFlow, PyTorch versions or specific library versions). |
| Experiment Setup | Yes | As the classifier layer is trained from scratch, we set its learning rate to be 10 times that of the other layers. We use mini-batch stochastic gradient descent (SGD) with 0.9 momentum and learning rate decaying strategy, and select learning rate between 10 5 and 10 2 by stepsize 10 1 2 . |