| CARVIEW |
Select Language
HTTP/2 301
server: GitHub.com
content-type: text/html
location: https://lvsn.github.io/associative-alignment/
access-control-allow-origin: *
expires: Tue, 30 Dec 2025 10:20:05 GMT
cache-control: max-age=600
x-proxy-cache: MISS
x-github-request-id: EE46:3FD64F:9F2F0B:B2BD8B:6953A4FC
accept-ranges: bytes
age: 0
date: Tue, 30 Dec 2025 10:10:05 GMT
via: 1.1 varnish
x-served-by: cache-bom-vanm7210066-BOM
x-cache: MISS
x-cache-hits: 0
x-timer: S1767089405.242411,VS0,VE203
vary: Accept-Encoding
x-fastly-request-id: 90d1dc6ea4ac017c0794a854c46245c0c9935beb
content-length: 162
HTTP/2 200
server: GitHub.com
content-type: text/html; charset=utf-8
last-modified: Thu, 11 Dec 2025 15:41:23 GMT
access-control-allow-origin: *
etag: W/"693ae623-263e"
expires: Tue, 30 Dec 2025 10:20:05 GMT
cache-control: max-age=600
content-encoding: gzip
x-proxy-cache: MISS
x-github-request-id: 5AC6:2DDCFF:9EBF7C:B24895:6953A4FB
accept-ranges: bytes
age: 0
date: Tue, 30 Dec 2025 10:10:05 GMT
via: 1.1 varnish
x-served-by: cache-bom-vanm7210066-BOM
x-cache: MISS
x-cache-hits: 0
x-timer: S1767089405.477750,VS0,VE215
vary: Accept-Encoding
x-fastly-request-id: fade3a059e57bcaf1f28e53df798b2708dc7e99e
content-length: 2958
Associative Alignment for Few-shot Image Classification
|
|
|
|
|
(as a spotlight presentation) |
| Few-shot image classification aims at training a model from only a few examples for each of the "novel" classes. This paper proposes the idea of associative alignment for leveraging part of the base data by aligning the novel training instances to the closely related ones in the base training set. This expands the size of the effective novel training set by adding extra "related base" instances to the few novel ones, thereby allowing a constructive fine-tuning. We propose two associative alignment strategies: 1) a metric-learning loss for minimizing the distance between related base samples and the centroid of novel instances in the feature space, and 2) a conditional adversarial alignment loss based on the Wasserstein distance. Experiments on four standard datasets and three backbones demonstrate that combining our centroid-based alignment loss results in absolute accuracy improvements of 4.4%, 1.2%, and 6.2% in 5-shot learning over the state of the art for object recognition, fine-grained classification, and cross-domain adaptation, respectively. |
 |
Arman Afrasiyabi, Jean-François Lalonde, Christian Gagné Associative Alignment for Few-shot Image Classification. (hosted on ArXiv) |
 |
Video (left: short, right: long)
