Papers with Code 아카이브

Self-Supervised Learning

태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.

하위 태스크
2
평가표
10
보존된 평가 기록
18

Self-Supervised Learning is proposed for utilizing unlabeled data with the success of supervised learning. Producing a dataset with good labels is expensive, while unlabeled data is being generated all the time. The motivation of Self-Supervised Learning is to make use of the large amount of unlabeled data. The main idea of Self-Supervised Learning is to generate the labels from unlabeled data, according to the structure or characteristics of the data itself, and then train on this unsupervised data in a supervised manner. Self-Supervised Learning is wildly used in representation learning to make a model learn the latent features of the data. This technique is often employed in computer vision, video processing and robot control.

Source: Self-supervised Point Set Local Descriptors for Point Cloud Registration

Image source: LeCun

데이터셋

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

Extended Agriculture-Vision

Extended Agriculture-Vision dataset comprises two parts: 1. An improved version of the Agriculture-Vision dataset, including full-field farmland imagery, encourages the exploration of geo-information on a larger scale. 2. Over three terabytes of high-resolution raw images across the US, aiming to inspire research in self-supervised learning in remote sensing and agriculture. Extended Agriculture-Vision identified 1200 fields from 2019-2020 growing seasons. Each image consists of RGB and Near-infrared (NIR) channels with resolutions as high as 10 cm per pixel.

OADAT

An experimental and synthetic (simulated) OA raw signals and reconstructed image domain datasets rendered with different experimental parameters and tomographic acquisition geometries. For detailed information, see github.com/berkanlafci/oadat.

Open Radar Datasets

A classification dataset of radar spectrograms in i "ground surveillance" setting recorded with the Open Radar Initiative. A dataset in a "ground surveillance" setting. The dataset has been collected with a stationary radar and targets moving in front of the radar. The dataset has been collected using both collaborative and non-collaborative targets.

Unified SSL Benchmark (USB)

The Unified SSL Benchmark (USB) consists of 15 diverse, challenging, and comprehensive tasks from CV, natural language processing (NLP), and audio processing (Audio) to evaluate self-supervised learning (SSL) methods. A modular and extensible codebase is open-sourced for fair evaluation on these SSL methods.

ApartmenTour

Contains a large number of online videos and subtitles. Source: Watch and Learn: Mapping Language and Noisy Real-world Videos with Self-supervision

diaforge-utc-r-0725

Dataset for our paper Disambiguation-Centric Finetuning Makes Enterprise Tool-Calling LLMs More Realistic and Less Risky which includes 5000 enterprise tools and the corresponding dialogues generated using DiaFORGE UTC data engine. The dataset is generated with the data generation engine. The engine simulates a user agent and an assistant agent in a dialogue, where the user agent has a persona and the assistant agent has access to a set of tools. Detailed information about the data generation process can be found in the paper. Each entry in the dataset contains: - seed: Gold tool for the generated dialogue. - user_persona: Persona of the simulated user agent. - messages: List of messages in the dialogue, where each message is a dictionary containing: - distractor_tools: List of distractor tools that are not the gold tool but are relevant to the dialogue. These tools are used by user agent to generate utterances that are hard to disambiguate from the gold tool. - retrieved_tools: List of tools retrieved by the assistant agent and used by the assistant agent in order to ask clarifying questions to the user agent.

GeoJEPAD

GeoJEPAD is a multimodal dataset combining OpenStreetMap (OSM) data (attributes and geometries) with high-resolution aerial imagery from diverse urban areas. Sourced from NAIP and OSM and then processed, tiled, and cropped. Geometries and relations represented as graphs with optional visibility edges. Motivation: Created to reduce biases introduced by traditional augmentation and sampling techniques, the dataset supports unbiased self-supervised multimodal geospatial learning. Potential Use Cases: - Self-supervised multimodal representation learning. - Semantic segmentation of aerial imagery. - Geospatial retrieval and urban analytics tasks. - Benchmarking multimodal fusion models like JEPA.

L1BSR

The Sentinel-2 satellite carries 12 CMOS detectors for the VNIR bands, with adjacent detectors having overlapping fields of view that result in overlapping regions in level-1 B (L1B) images. This dataset includes 3740 pairs of overlapping image crops extracted from two L1B products. Each crop has a height of around 400 pixels and a variable width that depends on the overlap width between detectors for RGBN bands, typically around 120-200 pixels. In addition to detector parallax, there is also cross-band parallax for each detector, resulting in shifts between bands. Pre-registration is performed for both cross-band and cross-detector parallax, with a precision of up to a few pixels (typically less than 10 pixels).

MAX-60K

The dataset for masked autoencoder for X-ray fluorescence (XRF) is a following development after the dataset (Chao et al., 2022). Besides the published XRF spectra-target measurements (CaCO3 and TOC) pairs of data, we further upload the XRF spectra in that project but without alignments of the target measurements here. As the first XRF large dataset compiled in a ML friendly format, we expect to kickoff more ML studies in the field of XRF and geology, especially DL studies. The investigated cores, which form the datast, are mostly retrieved across the high- to mid-latitude Northwest Pacific (37°N-52°N) and the Pacific sector of the Southern Ocean (53°S-63°S), with a water depth coverage from 1211 to 4853 m: <br> 1. Cruise SO264 in the subarctic Northwest Pacific with R/V SONNE in 2018<br> 2. Cruise PS97 in the central Drake Passage with RV Polarstern in 2016<br> 3. Cruise PS75 in the Pacific sector of the Southern Ocean in 2009/2010.<br> 4. Cruise KOMEX I and KOMEX II with R/V Akademik Lavrentyev in 1998 and cruise SO178 in 2004 in the Okhotsk Sea.

YFCC-CelebA

The scales of the data accessible through internet search engines can reach hundreds of millions, or even billions. The existence of such large weak-labeled databases has gained importance in the training of face recognition algorithms. Starting with the publicly available YFCC100M, we propose a weakly-labeled subset for multi-label face recognition for self-supervised methods. A 392K image subset of YFCC100M of 128x128 images was obtained by querying for the 40 facial attributes. We made this dataset publicly available.

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