# 5-way 5-shot Classification on TieredImageNet (Outlier Noise Robustness) Source: [5-way 5-shot Classification on TieredImageNet (Outlier Noise Robustness)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:33:58.350Z Evaluation of 5-way, 5-shot few-shot image classification performance on TieredImageNet, specifically measuring accuracy under varying levels of input outlier noise (0% to 60%). A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | 5-way 5-shot Classification | | Dataset (including any stated split) | TieredImageNet | | Metric | Accuracy (0% Noise) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 16 of 16 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Accuracy (0% Noise)](): Higher is better; unit: PERCENT. - [Accuracy (20% Noise)](): Higher is better; unit: PERCENT. - [Accuracy (40% Noise)](): Higher is better; unit: PERCENT. - [Accuracy (60% Noise)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy (0% Noise) | Result date | Report paper | | --- | --- | --- | --- | | Oracle | 71.42 | Not recorded | 68fa6d1b1e209131cea37c7f | | Vanilla ProtoNet | 71.42 | Not recorded | 68fa6d1b1e209131cea37c7f | | Baseline++ | 71.29 | Not recorded | 68fa6d1b1e209131cea37c7f | | RNNP | 71.28 | Not recorded | 68fa6d1b1e209131cea37c7f | | Spatial median prototypes (Median) | 71.28 | Not recorded | 68fa6d1b1e209131cea37c7f | | Similarity weighted prototypes (Euclidean) | 71.28 | Not recorded | 68fa6d1b1e209131cea37c7f | | Similarity weighted prototypes (Absolute) | 71.17 | Not recorded | 68fa6d1b1e209131cea37c7f | | TraNFS-3 | 71.13 | Not recorded | 68fa6d1b1e209131cea37c7f | | TraNFS-2 | 70.83 | Not recorded | 68fa6d1b1e209131cea37c7f | | Similarity weighted prototypes (Cosine) | 70.79 | Not recorded | 68fa6d1b1e209131cea37c7f | | Linear Classifier | 69.6 | Not recorded | 68fa6d1b1e209131cea37c7f | | Matching Networks | 64.99 | Not recorded | 68fa6d1b1e209131cea37c7f | | MAML | 63.9 | Not recorded | 68fa6d1b1e209131cea37c7f | | Nearest k = 5 | 59.25 | Not recorded | 68fa6d1b1e209131cea37c7f | | Nearest k = 1 | 58.89 | Not recorded | 68fa6d1b1e209131cea37c7f | | Nearest k = 3 | 58.38 | Not recorded | 68fa6d1b1e209131cea37c7f | ### Oracle Accuracy (0% Noise): 71.42. Result ID: 698f8f023a6c314849fa193f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Protocol":"5-way 5-shot"}. This record does not assert independent reproduction. ### Vanilla ProtoNet Accuracy (0% Noise): 71.42. Result ID: 698f8f023a6c314849fa1946. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### Baseline++ Accuracy (0% Noise): 71.29. Result ID: 698f8f023a6c314849fa1947. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### RNNP Accuracy (0% Noise): 71.28. Result ID: 698f8f023a6c314849fa1948. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### Spatial median prototypes (Median) Accuracy (0% Noise): 71.28. Result ID: 698f8f023a6c314849fa1949. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Aggregation method":"Median","Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### Similarity weighted prototypes (Euclidean) Accuracy (0% Noise): 71.28. Result ID: 698f8f023a6c314849fa194b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Aggregation method":"Euclidean distance weighting","Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### Similarity weighted prototypes (Absolute) Accuracy (0% Noise): 71.17. Result ID: 698f8f023a6c314849fa194a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Aggregation method":"Absolute distance weighting","Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### TraNFS-3 Accuracy (0% Noise): 71.13. Result ID: 698f8f023a6c314849fa194e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Number of Layers":3,"Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### TraNFS-2 Accuracy (0% Noise): 70.83. Result ID: 698f8f023a6c314849fa194d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Number of Layers":2,"Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### Similarity weighted prototypes (Cosine) Accuracy (0% Noise): 70.79. Result ID: 698f8f023a6c314849fa194c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Aggregation method":"Cosine similarity weighting","Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### Linear Classifier Accuracy (0% Noise): 69.6. Result ID: 698f8f023a6c314849fa1943. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### Matching Networks Accuracy (0% Noise): 64.99. Result ID: 698f8f023a6c314849fa1944. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### MAML Accuracy (0% Noise): 63.9. Result ID: 698f8f023a6c314849fa1945. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### Nearest k = 5 Accuracy (0% Noise): 59.25. Result ID: 698f8f023a6c314849fa1942. [Reporting paper]() Recorded evaluation configuration: {"k":5,"Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### Nearest k = 1 Accuracy (0% Noise): 58.89. Result ID: 698f8f023a6c314849fa1940. [Reporting paper]() Recorded evaluation configuration: {"k":1,"Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. ### Nearest k = 3 Accuracy (0% Noise): 58.38. Result ID: 698f8f023a6c314849fa1941. [Reporting paper]() Recorded evaluation configuration: {"k":3,"Backbone":"4-layer convolutional"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.