Probabilistic two-stage detection

Two-stage object detectors that use class-agnostic one-stage detectors as the proposal network.

Github: https://github.com/xingyizhou/CenterNet2

Paper: https://arxiv.org/abs/2103.07461v1


@ai_machinelearning_big_data

🔷 PyTorch Geometric Temporal is a temporal (dynamic) extension library for PyTorch Geometric.

Github: https://github.com/benedekrozemberczki/pytorch_geometric_temporal

Paper: https://dl.acm.org/doi/10.1145/3308560.3316581

Dataset: https://pytorch-geometric-temporal.readthedocs.io/en/latest/notes/introduction.html#discrete-time-datasets

@ai_machinelearning_big_data

Large Scale Image Completion via Co-Modulated Generative Adversarial Networks, ICLR 2021 (Spotlight)

Gitub: https://github.com/zsyzzsoft/co-mod-gan

Paper: https://openreview.net/pdf?id=sSjqmfsk95O

Demo: http://comodgan.ml/

@ai_machinelearning_big_data

Contrastive Learning of Musical Representations

Github: https://github.com/spijkervet/CLMR

Paper: https://arxiv.org/abs/2103.09410

@ai_machinelearning_big_data

🔬 The Best ML Frameworks & Extensions for Scikit-learn

https://neptune.ai/blog/the-best-ml-framework-extensions-for-scikit-learn

@ai_machinelearning_big_data

The last article in uplift modeling tutorials by MTS data scientists: “Metrics in uplift modeling”📈

Uplift models are used to promote products and create advertising segments 🤑

📗 Habr: https://habr.com/ru/company/ru_mts/blog/538934/
💻 Code: https://nbviewer.jupyter.org/github/maks-sh/scikit-uplift/blob/master/notebooks/uplift_metrics_tutorial.ipynb

The GENRE (Generarive ENtity REtrieval) system as presented in Autoregressive Entity Retrieval implemented in pytorch.

Github: https://github.com/facebookresearch/GENRE

Paper: https://arxiv.org/abs/2103.12528v1

Dataset: https://github.com/facebookresearch/GENRE/blob/main/scripts/download_all_datasets.sh

@ai_machinelearning_big_data

Каждый день мы загружаем и передаем все больше контента. Нагрузка на базовые станции операторов растет, но вместе с ней растет и качество связи. Как это получается? И какую роль в этом процессе играет big data?

Ответы найдете в видео «Как нейросети помогают оставаться на связи», где Тимур и Вика, датасайентисты из МегаФона, рассказывают, почему модернизация базовых станций – это задача из области data science, как работает модель, на основании рекомендаций которой принимаются решения о необходимости «прокачки» той или иной вышки, и остается ли в этом процессе место для человека.

Спойлер: место для человека остается. Например, в МегаФоне в команду по работе с большими данными ищут Team Lead, Senior data scientist, инженера SQL и других специалистов.

Towards Ultra-Resolution Neural Style Transfer via Thumbnail Instance Normalization

Github: https://github.com/czczup/URST

Paper: https://arxiv.org/abs/2103.11784

Dataset: https://cocodataset.org/#download

@ai_machinelearning_big_data

🤗 ⚡️Multilingual CLIP with Huggingface + PyTorch Lightning

https://sachinruk.github.io/blog/pytorch/pytorch%20lightning/loss%20function/gpu/2021/03/07/CLIP.html#Data

CLIP: Connecting Text and Images: https://openai.com/blog/clip/


@ai_machinelearning_big_data

Swin Transformer - PyTorch

Code: https://github.com/berniwal/swin-transformer-pytorch

Github: https://github.com/microsoft/Swin-Transformer

Paper: https://arxiv.org/abs/2103.14030v1

@ai_machinelearning_big_data

UniverseNet

UniverseNet is the state-of-the-art detector that can be trained in 24 epochs.

Github: https://github.com/shinya7y/UniverseNet

Paper: https://arxiv.org/abs/2103.14027v1

@ai_machinelearning_big_data

TransT - Transformer Tracking

Github: https://github.com/chenxin-dlut/TransT

Paper: https://arxiv.org/abs/2103.15436

Model: https://drive.google.com/drive/folders/1GVQV1GoW-ttDJRRqaVAtLUtubtgLhWCE

@ai_machinelearning_big_data

Hyperbolic Graph Convolutional Auto-Encoders

Github: https://github.com/junhocho/HGCAE

Paper: https://arxiv.org/abs/2103.16046

@ai_machinelearning_big_data

⚡️ NeX: Real-time View Synthesis with Neural Basis Expansion

Github: https://github.com/nex-mpi/nex-code

Real-time View Synthesis with Python Code: https://analyticsindiamag.com/what-is-nex-guide-to-real-time-view-synthesis-with-python-code/

Project: https://nex-mpi.github.io/

Paper: https://arxiv.org/pdf/2103.05606.pdf

Video: https://www.youtube.com/watch?v=HyfkF7Z-ddA&ab_channel=SupasornSuwajanakorn

@ai_machinelearning_big_data

EasyMocap is an open-source toolbox for markerless human motion capture from RGB videos

Github: https://github.com/zju3dv/EasyMocap

Paper: https://arxiv.org/abs/2104.00340v1

@ai_machinelearning_big_data

👁 Learning Spatio-Temporal Transformer for Visual Tracking

Github: https://github.com/researchmm/Stark

Paper: https://arxiv.org/abs/2103.17154v1

@ai_machinelearning_big_data

Audio-Visual Event Localization in Unconstrained Videos

https://sites.google.com/view/audiovisualresearch

Github: https://github.com/YapengTian/AVE-ECCV18

Paper: https://arxiv.org/abs/1803.08842

@ai_machinelearning_big_data

Timers and Such: A Practical Benchmark for Spoken Language Understanding with Numbers

Github: https://github.com/speechbrain/speechbrain/tree/develop/recipes/timers-and-such

Dataset: https://zenodo.org/record/4623772#.YG08zhQzZs9

Paper: https://arxiv.org/abs/2104.01604v1

@ai_machinelearning_big_data

Regularizing Generative Adversarial Networks under Limited Data

Github: https://github.com/google/lecam-gan

Paper: https://arxiv.org/abs/2104.03310

https://hytseng0509.github.io/lecam-gan/

@ai_machinelearning_big_data

📗 New updates: Interpretable Machine Learning

A Guide for Making Black Box Models Explainable.

https://christophm.github.io/interpretable-ml-book/

@ai_machinelearning_big_data