Data Phoenix Digest - ISSUE 29

NEWS

What's new this week?

AI adoption accelerates. AI-powered drive-thrus. NATO's focus on AI. AI's progress in tackling racial inequality, healthcare issues, enzyme research, and supply chain disruptions.

Funding News

  • Mage, a developer of an AI tool for product developers to build and integrate AI into apps, raises $6.3M in seed funding led by Gradient Ventures.
  • Hex, a collaborative data workspace for data scientists, brings in $16M in Series A funding led by Tomasz Tunguz from Redpoint Ventures.
  • Abacus.ai, a company building an AI platform that simplifies ML model creation, raises $50M in Series C led by Tiger Global; announced support of CV use cases.
  • Resistant AI, a developer of AI-powered anti-fraud solution, closes $16.6M in Series A funding led by GV (Google Ventures) and existing investors.
  • Deci, a Tel Aviv-based startup for building usable models to run AI algorithms, picks up a Series A of $21M. Insight Partners and existing investors lead the round.

ARTICLES

Introduction to Distributed Training in PyTorch
This tutorial will introduce you to the basics of distributed training in PyTorch. Note that this is the last of article in a 3-part tutorial on intermediate PyTorch techniques for CV/DL practitioners.

Deploying Serverless spaCy Transformer Model with AWS Lambda
In this step-by-step guide, you'll learn how to deploy NER transformer model serverless with Huggingface and AWS Lambda to run predictions. Four major steps.

Reflections on Foundation Models
In this blog post, The Gradient team talk through why foundation models are important and clarify several points in relation to the community response to the original report.

Predicting Gene Expression with AI
In this article, DeepMind team go deep into their new Enformer architecture that advances genetic research by improving the ability to predict how DNA sequence influences gene expression.

State of AI Report 2021
This year’s State of AI Report points to the real-world performance breakthroughs in NLP, CV, and biology. Research into AI safety lags behind its rapid commercial, civil, and military deployment.

CLOOB: Modern Hopfield Networks with InfoLOOB Outperform CLIP
This article is a comprehensive, detailed overview and explanation of the paper: “CLOOB: Modern Hopfield Networks with InfoLOOB Outperform CLIP”. Learn more!

PAPERS

Applications and Techniques for Fast Machine Learning in Science
This community review report discusses applications and techniques for fast ML in science based on two workshops held by the Fast ML for Science community. Make sure to have a look!

DocTr: Document Image Transformer for Geometric Unwarping and Illumination Correction
Hao Feng et al. propose a new framework, called Document Image Transformer (DocTr), to address the issue of geometry and illumination distortion of the document images.

Resolution-robust Large Mask Inpainting with Fourier Convolutions
In this paper, the researchers propose a new method called large mask inpainting (LaMa) to solve the issue of an effective receptive field in both the inpainting network and the loss function.

SciCap: Generating Captions for Scientific Figures
In this paper, the researchers propose an end-to-end neural framework, called SciCap, to automatically generate informative, high-quality captions for scientific figures.

CIPS-3D: A 3D-Aware Generator of GANs Based on Conditionally-Independent Pixel Synthesis
In this paper, Peng Zhou et al. present CIPS-3D, a style-based, 3D-aware generator that is composed of a shallow NeRF network and a deep implicit neural representation (INR) network.

SgSum: Transforming Multi-document Summarization into Sub-graph Selection
The researchers propose a novel MDS framework to formulate the MDS task as a sub-graph selection problem. The architecture has strong transfer ability from single to multi-document input

EVENTS

GTC [NVIDIA Conference]
NVIDIA GTC is more than a must-attend AI conference for developers (Nov. 8-11). Don’t miss a keynote from Jensen Huang, plus sessions and talks with luminaries from around the world.

COURSES

Knowledge Graphs Course 2021 [in Russian]
Graph Representation Learning (GRL) is an extremely popular area of research. This course aims to close the gaps Russian speakers may have in the topic. 9 lectures in total.

Your Second RecSys Course [in Russian]
Your Second RecSys course is the continuation of Your First RecSys, a course on designing and building recommendation system. Check out the first course, too.

VIDEOS

IGDL 2021 Israeli Geometric Deep Learning Workshop
Listen to three exclusive keynotes and 15 short talks delivered at the second Israeli workshop on Geometric Deep Learning. Major topics: DL on Non-Euclidian domains and its applications.

JOBS

Looking to feature your open positions in the digest? Kindly reach out to us at editor@dataphoenix.info for details. We'll be proud to help your business thrive!