Prototype Memory for Large-scale Face Representation Learning
- URL: http://arxiv.org/abs/2105.02103v1
- Date: Wed, 5 May 2021 15:08:34 GMT
- Title: Prototype Memory for Large-scale Face Representation Learning
- Authors: Evgeny Smirnov, Nikita Garaev, Vasiliy Galyuk
- Abstract summary: Softmax-based approach is not suitable for datasets with millions of persons.
We propose a novel face representation learning model called Prototype Memory.
We prove the effectiveness of the proposed model by extensive experiments on popular face recognition benchmarks.
- Score: 0.5524804393257919
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: Face representation learning using datasets with massive number of identities
requires appropriate training methods. Softmax-based approach, currently the
state-of-the-art in face recognition, in its usual "full softmax" form is not
suitable for datasets with millions of persons. Several methods, based on the
"sampled softmax" approach, were proposed to remove this limitation. These
methods, however, have a set of disadvantages. One of them is a problem of
"prototype obsolescence": classifier weights (prototypes) of the rarely sampled
classes, receive too scarce gradients and become outdated and detached from the
current encoder state, resulting in an incorrect training signals. This problem
is especially serious in ultra-large-scale datasets. In this paper, we propose
a novel face representation learning model called Prototype Memory, which
alleviates this problem and allows training on a dataset of any size. Prototype
Memory consists of the limited-size memory module for storing recent class
prototypes and employs a set of algorithms to update it in appropriate way. New
class prototypes are generated on the fly using exemplar embeddings in the
current mini-batch. These prototypes are enqueued to the memory and used in a
role of classifier weights for usual softmax classification-based training. To
prevent obsolescence and keep the memory in close connection with encoder,
prototypes are regularly refreshed, and oldest ones are dequeued and disposed.
Prototype Memory is computationally efficient and independent of dataset size.
It can be used with various loss functions, hard example mining algorithms and
encoder architectures. We prove the effectiveness of the proposed model by
extensive experiments on popular face recognition benchmarks.
Related papers
- Unicom: Universal and Compact Representation Learning for Image
Retrieval [65.96296089560421]
We cluster the large-scale LAION400M into one million pseudo classes based on the joint textual and visual features extracted by the CLIP model.
To alleviate such conflict, we randomly select partial inter-class prototypes to construct the margin-based softmax loss.
Our method significantly outperforms state-of-the-art unsupervised and supervised image retrieval approaches on multiple benchmarks.
arXiv Detail & Related papers (2023-04-12T14:25:52Z) - Learning Large Scale Sparse Models [6.428186644949941]
We consider learning sparse models in large scale settings, where the number of samples and the feature dimension can grow as large as millions or billions.
We propose to learn sparse models such as Lasso in an online manner where in each, only one randomly chosen sample is revealed to update a sparse gradient.
Thereby, the memory cost is independent of the sample size and gradient evaluation for one sample is efficient.
arXiv Detail & Related papers (2023-01-26T06:29:49Z) - Incremental Online Learning Algorithms Comparison for Gesture and Visual
Smart Sensors [68.8204255655161]
This paper compares four state-of-the-art algorithms in two real applications: gesture recognition based on accelerometer data and image classification.
Our results confirm these systems' reliability and the feasibility of deploying them in tiny-memory MCUs.
arXiv Detail & Related papers (2022-09-01T17:05:20Z) - Dynamic Prototype Mask for Occluded Person Re-Identification [88.7782299372656]
Existing methods mainly address this issue by employing body clues provided by an extra network to distinguish the visible part.
We propose a novel Dynamic Prototype Mask (DPM) based on two self-evident prior knowledge.
Under this condition, the occluded representation could be well aligned in a selected subspace spontaneously.
arXiv Detail & Related papers (2022-07-19T03:31:13Z) - A Model or 603 Exemplars: Towards Memory-Efficient Class-Incremental
Learning [56.450090618578]
Class-Incremental Learning (CIL) aims to train a model with limited memory size to meet this requirement.
We show that when counting the model size into the total budget and comparing methods with aligned memory size, saving models do not consistently work.
We propose a simple yet effective baseline, denoted as MEMO for Memory-efficient Expandable MOdel.
arXiv Detail & Related papers (2022-05-26T08:24:01Z) - Rethinking Semantic Segmentation: A Prototype View [126.59244185849838]
We present a nonparametric semantic segmentation model based on non-learnable prototypes.
Our framework yields compelling results over several datasets.
We expect this work will provoke a rethink of the current de facto semantic segmentation model design.
arXiv Detail & Related papers (2022-03-28T21:15:32Z) - SSSE: Efficiently Erasing Samples from Trained Machine Learning Models [103.43466657962242]
We propose an efficient and effective algorithm, SSSE, for samples erasure.
In certain cases SSSE can erase samples almost as well as the optimal, yet impractical, gold standard of training a new model from scratch with only the permitted data.
arXiv Detail & Related papers (2021-07-08T14:17:24Z) - Detecting Anomalies in Semantic Segmentation with Prototypes [23.999211737485812]
We propose to address anomaly segmentation through prototype learning.
Our approach achieves the new state of the art, with a significant margin over previous works.
arXiv Detail & Related papers (2021-06-01T13:22:33Z) - One Line To Rule Them All: Generating LO-Shot Soft-Label Prototypes [0.0]
Prototype generation methods aim to create a small set of synthetic observations that accurately represent a training dataset.
assigning soft labels to prototypes can allow increasingly small sets of prototypes to accurately represent the original training dataset.
We propose a novel, modular method for generating soft-label lines that still maintains representational accuracy even when there are fewer prototypes than the number of classes in the data.
arXiv Detail & Related papers (2021-02-15T20:21:29Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.