TADAM: Task dependent adaptive metric for improved few-shot learning

Authors: Boris Oreshkin, Pau Rodríguez López, Alexandre Lacoste

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Table 1 presents our key result in the context of existing state-of-the-art. The five first rows show approaches that use the same feature extractor as [33]...As it can be seen, the proposed algorithm significantly improves over the existing state-of-the-art results on the mini-Imagenet dataset. In the rest of the section we address the following research questions: (i) can metric scaling improve few-shot classification results? (Sections 3.2 and 3.4), (ii) what are the contributions of each components of our proposed architecture? (Section 3.4), (iii) can task conditioning improve few-shot classification results and how important it is at different feature extractor depths? (Sections 3.3 and 3.4), and (iv) can auxiliary classification task co-training improve accuracy on the few-shot classification task? (Section 3.4).
Researcher Affiliation Collaboration Boris N. Oreshkin Element AI boris@elementai.com Pau Rodriguez Element AI, CVC-UAB pau.rodriguez@elementai.com Alexandre Lacoste Element AI allac@elementai.com
Pseudocode No No explicit pseudocode or algorithm blocks were found in the paper. The methodology is described in prose and through architectural diagrams.
Open Source Code No The paper does not provide any explicit statements about open-source code release or links to a code repository for the described methodology.
Open Datasets Yes mini-Imagenet. The mini-Imagenet dataset was proposed by Vinyals et al. [33].
Dataset Splits Yes To perform meta-validation and meta-test on unseen tasks (and classes), we isolate 16 and 20 classes from the original set of 100, leaving 64 classes for the training tasks. We use exactly the same train/validation/test split as the one suggested by Ravi and Larochelle [22].
Hardware Specification No The paper does not provide specific details on the hardware used for running the experiments, such as GPU models, CPU types, or memory specifications.
Software Dependencies No The paper does not provide specific version numbers for software dependencies or libraries used in the experiments.
Experiment Setup Yes The auxiliary task is sampled with a probability that is annealed over episodes. We annealed it using an exponential decay schedule of the form 0.9b20t/T c, where T is the total number of training episodes, t is episode index. The initial auxiliary task selection probability was cross-validated to be 0.9 and the number of decay steps was chosen to be 20.