@data_analysis_ml - Data analysis blog

🎆 Optimizing Relevance Maps of Vision Transformers Improves Robustness

This code allows to finetune the explainability maps of Vision Transformers to enhance robustness.

Github: https://github.com/hila-chefer/robustvit

Colab: https://colab.research.google.com/github/hila-chefer/RobustViT/blob/master/RobustViT.ipynb

Paper: https://arxiv.org/abs/2206.01161

Dataset: https://github.com/UnsupervisedSemanticSegmentation/ImageNet-S

@ai_machinelearning_big_data

UniSRec

The proposed approach utilizes the associated denoscription text of items to learn transferable representations across different recommendation scenarios.

Github: https://github.com/rucaibox/unisrec

Paper: https://arxiv.org/abs/2206.05941v1

Google Drive: https://drive.google.com/drive/folders/1Uik0fMk4oquV_bS9lXTZuExAYbIDkEMW?usp=sharing

🔘 Squeezeformer: An Efficient Transformer for Automatic Speech Recognition

Github: https://github.com/kssteven418/squeezeformer

Paper: https://arxiv.org/abs/2206.00888v1

Dataset: https://paperswithcode.com/dataset/librispeech

@ai_machinelearning_big_data

OntoMerger: An Ontology Integration Library for Deduplicating and Connecting Knowledge Graph Nodes

OntoMerger is an ontology alignment library for deduplicating knowledge graph nodes

Github: https://github.com/astrazeneca/onto_merger

Paper: https://arxiv.org/abs/2206.02238v1

Documentation: https://ontomerger.readthedocs.io/

@ai_machinelearning_big_data

👁‍🗨 CVNets: A library for training computer vision networks

Improved model, MobileViTv2, is state-of-the-art on several mobile vision tasks, including ImageNet object classification and MS-COCO object detection.

Github: https://github.com/apple/ml-cvnets

Examples: https://github.com/apple/ml-cvnets/blob/main/docs/source/en/models

Paper: https://arxiv.org/abs/2206.02680v1

Dataset: https://paperswithcode.com/dataset/coco

@ai_machinelearning_big_data

🔥 EG3D: Efficient Geometry-aware 3D Generative Adversarial Networks by Nvidia

Expressive hybrid explicit-implicit network architecture that, together with other design choices, synthesizes not only high-resolution multi-view-consistent images in real time but also produces high-quality 3D geometry.

Github: https://github.com/NVlabs/eg3d

Project: https://nvlabs.github.io/eg3d/

Video: https://www.youtube.com/watch?v=cXxEwI7QbKg&feature=emb_logo&ab_channel=StanfordComputationalImagingLab

Paper: https://nvlabs.github.io/eg3d/media/eg3d.pdf

@ai_machinelearning_big_data

🔦 Featurized Query R-CNN

Featurized object queries predicted by a query generation network in the well-established Faster R-CNN framework and develop a Featurized Query R-CN

Github: https://github.com/hustvl/featurized-queryrcnn

Paper: https://arxiv.org/abs/2206.06258v1

Dataset: https://paperswithcode.com/dataset/crowdhuman

@ai_machinelearning_big_data

Can CNNs Be More Robust Than Transformers?

CNN architectures without any attention-like operations that is as robust as, or even more robust than, Transformers.

Github: https://github.com/ucsc-vlaa/robustcnn

Paper: https://arxiv.org/abs/2206.03452v1

Dataset: https://paperswithcode.com/dataset/imagenet-r

@ai_machinelearning_big_data

@itchannels_telegram - data science, machine learning useful channels

😲 LIVE- Towards Layer-wise Image Vectorization (CVPR 2022 Oral)

New method to progressively generate a SVG that fits the raster image in a layer-wise fashion.

Github: https://github.com/picsart-ai-research/live-layerwise-image-vectorization

Project: https://ma-xu.github.io/LIVE/

Paper: https://arxiv.org/pdf/2206.04655v1.pdf

Colab: https://colab.research.google.com/drive/1s108WmqSVH9MILOjSAu29QyAEjExOWAP?usp=sharing

@ai_machinelearning_big_data

🪁 Age prediction of a speaker's voice

https://miykael.github.io/blog/2022/audio_eda_and_modeling/

@ai_machinelearning_big_data

🔖 DoWhy | An end-to-end library for causal inference

"DoWhy" is a Python library that aims to spark causal thinking and analysis.

Github: https://github.com/py-why/dowhy

Docs: https://py-why.github.io/dowhy/

Paper: https://arxiv.org/abs/2206.06821v1

Video: https://note.microsoft.com/MSR-Webinar-DoWhy-Library-Registration-On-Demand.html

@ai_machinelearning_big_data

🌩 Object Structural Points Representation for Graph-based Semantic Monocular Localization and Mapping

PyGOD is a Python library for graph outlier detection (anomaly detection).

Github: https://github.com/pygod-team/pygod

Dataset : https://paperswithcode.com/dataset/ogb

Paper: https://arxiv.org/abs/2206.10071v1

🖊 StrengthNet

Implementation of "StrengthNet: Deep Learning-based Emotion Strength Assessment for Emotional Speech Synthesis"

Github: https://github.com/ttslr/strengthnet

Paper: https://arxiv.org/abs/2110.03156

MOSNet: https://github.com/lochenchou/MOSNet

@ai_machinelearning_big_data

✔️ Pythae: Unifying Generative Autoencoders in Python -- A Benchmarking Use Case

This library implements some of the most common (Variational) Autoencoder models.

Github: https://github.com/clementchadebec/benchmark_VAE

Paper: https://arxiv.org/abs/2206.08309v1

Dataset: https://paperswithcode.com/dataset/celeba

@ai_machinelearning_big_data

💻 DaisyRec 2.0: Benchmarking Recommendation for Rigorous Evaluation

DaisyRec-v2.0 is a Python toolkit developed for benchmarking top-N recommendation task.

Github: https://github.com/recsys-benchmark/daisyrec-v2.0

Command Generator : http://daisyrecguicommandgenerator.pythonanywhere.com/

Paper: https://arxiv.org/abs/2206.10848v1

Tutorial: https://github.com/recsys-benchmark/DaisyRec-v2.0/blob/main/DaisyRec-v2.0-Tutorial.ipynb

🔊 SoundSpaces 2.0: A Simulation Platform for Visual-Acoustic Learning

We introduce SoundSpaces 2.0, a platform for on-the-fly geometry-based audio rendering for 3D environments.

Github: https://github.com/facebookresearch/sound-spaces

Paper: https://arxiv.org/abs/2206.08312v1

Dataset: https://paperswithcode.com/dataset/librispeech

@ai_machinelearning_big_data

📡 NU-Wave — Official PyTorch Implementation

Github: https://github.com/mindslab-ai/nuwave

Paper: https://arxiv.org/abs/2206.08545v1

Dataset: https://datashare.ed.ac.uk/handle/10283/3443

@ai_machinelearning_big_data

Frequency Dynamic Convolution-Recurrent Neural Network (FDY-CRNN) for Sound Event Detection

Frequency Dynamic Convolution applied kernel that adapts to each freqeuncy bin of input, in order to remove tranlation equivariance of 2D convolution along the frequency axis.

Github: https://github.com/frednam93/FDY-SED

Paper: https://arxiv.org/abs/2206.11645v1

Dataset: https://paperswithcode.com/dataset/desed

@ai_machinelearning_big_data

♦️ Color engineering for special images

How to improve color encoding of unnatural images.

Article
Dataset

@ai_machinelearning_big_data