Papers with Code 아카이브

Continual Learning

태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.

하위 태스크
5
평가표
33
보존된 평가 기록
133

Continual Learning (also known as Incremental Learning, Life-long Learning) is a concept to learn a model for a large number of tasks sequentially without forgetting knowledge obtained from the preceding tasks, where the data in the old tasks are not available anymore during training new ones.
If not mentioned, the benchmarks here are Task-CL, where task-id is provided on validation.

Source:
Continual Learning by Asymmetric Loss Approximation with Single-Side Overestimation
Three scenarios for continual learning
Lifelong Machine Learning
Continual lifelong learning with neural networks: A review

데이터셋

출처에 같은 태스크 이름으로 연결된 데이터셋입니다. 동명 태스크가 포함될 수 있으므로, 비교할 때는 평가 결과의 조건도 확인하세요.

ConCon Dataset

ConCon: Continually Confounded Dataset is a confounded visual dataset for continual learning. Files contained: case_disjoint, case_strict and unconfounded consist of splits for training (train), validation (val), testing (test). Each split within case_disjoint and case_strict contains 3 tasks (t0, t1, t2) whereas there is 1 task (t0) within unconfounded. Each task contains folders for postive(1) and negative(0) images.

CRL-Person

Provides two large-scale multi-step benchmarks for biometric identification, where the visual appearance of different classes are highly relevant. Source: Continual Representation Learning for Biometric Identification

HOWS

HOWS-CL-25 (Household Objects Within Simulation dataset for Continual Learning) is a synthetic dataset especially designed for object classification on mobile robots operating in a changing environment (like a household), where it is important to learn new, never seen objects on the fly. This dataset can also be used for other learning use-cases, like instance segmentation or depth estimation. Or where household objects or continual learning are of interest. Our dataset contains 150,795 unique synthetic images using 25 different household categories with 925 3D models in total. For each of those categories, we generated about 6000 RGB images. In addition, we also provide a corresponding depth, segmentation, and normal image. The dataset was created with BlenderProc [Denninger et al. (2019)], a procedural pipeline to generate images for deep learning. This tool created a virtual room with randomly textured floors, walls, and a light source with randomly chosen light intensity and color. After that, a 3D model is placed in the resulting room. This object gets customized by randomly assigning materials, including different textures, to achieve a diverse dataset. Moreover, each object might be deformed with a random displacement texture. We use 774 3D models from the ShapeNet dataset [A. X. Chang et al. (2015)] and the other models from various internet sites. Please note that we had to manually fix and filter most of the models with Blender before using them in the pipeline! For continual learning (CL), we provide two different loading schemes: - Five sequences with five categories each - Twelve sequences with three categories in the first and two in the other sequences. In addition to the RGB, depth, segmentation, and normal images, we also provide the calculated features of the RGB images (by ResNet50) as used in our RECALL paper. In those two loading schemes, ten percent of the images are used for validation, where we ensure that an object instance is either in the training or the validation set, not in both. This avoids learning to recognize certain instances by heart. We recommend using those loading schemes to compare your approach with others. For further information and code examples, please have a look at our website: https://github.com/DLR-RM/RECALL.

idsprites

Easily generate simple continual learning benchmarks. Inspired by dSprites. > It allows for procedurally generating a virtually infinite progression of random two-dimensional shapes. [...] we generate each unique shape in every combination of orientation, scale, and position. Most importantly, by providing the ground truth values of individual factors of variation (FoVs)

SPOT-10

The SPOTS-10 dataset is an extensive collection of grayscale images showcasing diverse patterns found in ten animal species. Specifically, SPOTS-10 contains 50,000 32 x 32 grayscale images, divided into ten categories, with 5,000 images per category. The training set comprises 40,000 images, while the test set contains 10,000 images. The SPOTS-10 dataset is freely available on the project GitHub page: https://github.com/amotica/spots-10 by cloning the repository.

TiROD

Dataset to benchmark Continual Learning for Object Detection in a Tiny Robotics settings. * 10 Continual learning tasks * 5 environments (1 indoor and 4 outdoor) * 2 illumination conditions * 14 object categories Dataset Website Dataset video Code Paper

출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.