Incremental Few-Shot Learning with Attention Attractor Networks

Authors: Mengye Ren, Renjie Liao, Ethan Fetaya, Richard Zemel

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We show empirically that our proposed method can produce state-of-the-art results in incremental few-shot learning on mini-Image Net [36] and tiered-Image Net [29] tasks.
Researcher Affiliation Collaboration Mengye Ren1,2,3, Renjie Liao1,2,3, Ethan Fetaya1,2, Richard S. Zemel1,2 1University of Toronto, 2Vector Institute, 3Uber ATG
Pseudocode Yes Algorithm 1 Meta Learning for Incremental Few-Shot Learning
Open Source Code Yes Code released at: https://github.com/renmengye/inc-few-shot-attractor-public
Open Datasets Yes We experiment on two few-shot classification datasets, mini-Image Net and tiered-Image Net. Both are subsets of Image Net [30]... mini-Image Net Proposed by [36]... tiered-Image Net Proposed by [29]
Dataset Splits Yes mini-Image Net Proposed by [36], mini-Image Net contains 100 object classes and 60,000 images. We used the splits proposed by [27], where training, validation, and testing have 64, 16 and 20 classes respectively.
Hardware Specification No The paper does not provide specific details about the hardware used for experiments, such as GPU models, CPU types, or memory specifications.
Software Dependencies No The paper mentions software components like ResNet, L-BFGS, and ADAM optimizer, but it does not specify exact version numbers for any libraries or frameworks (e.g., PyTorch, TensorFlow versions).
Experiment Setup Yes We use a standard Res Net backbone [11]... We use L-BFGS [43] to solve the inner loop of our models... We use the ADAM [14] optimizer for meta-learning with a learning rate of 1e-3, which decays by a factor of 10 after 4,000 steps, for a total of 8,000 steps. We fix recurrent backpropagation to 20 iterations and ϵ = 0.1.