🤖✨ ML, Physics and Robotics
Free news digest curated by the industry insiders
📝 Editorial
Robotics is one of the quintessential use cases for machine learning (ML) but also one of the most difficult ones to implement. Robotic applications need to combine hardware and ML models in very complex ways and are subjected to interactions with real-world environments, which are very hard to model. Object interactions, texture, environmental dynamics, advanced geometry are common elements that need to be present in physic simulation models. Not surprisingly, physics simulation is one of the hardest things to get right in ML solutions. The absence of robust simulators remains one of the biggest obstacles to making robotics research mainstream. Many of the physic simulation stacks remain closed-source and based on proprietary solutions. However, the ML industry has been making steady progress in this area.
A few days ago, DeepMind announced that it was acquiring and open-sourcing MuJoCo ( Multi-Joint dynamics with Contact), one of the most popular frameworks for robotics research. MuJoCo is a general-purpose physics engine that enables robust simulation of articulated physical structures and environments. Open-sourcing MuJoCo and combining with DeepMind’s advanced ML models in areas such as reinforcement learning could really accelerate robotics research and applying ML in physics in general.
🔺🔻TheSequence Scope – our Sunday edition with the industry’s development overview – is free. To receive high-quality content about the most relevant developments in the ML world every Tuesday and Thursday, please subscribe to TheSequence Edge 🔺🔻
🗓 Next week in TheSequence Edge:
Edge#135: we discuss Self-Supervised Learning for Computer Vision; we explore SEER, one of the most powerful SSL models for computer vision ever built; we cover Hugging Face library for computer vision, including such models as Vision Transformer, VisualBERT, and DeIT.
Edge#136: deep dive about Kili Technology, an AI training platform for Data-Centric companies.
Now, let’s review the most important developments in the AI industry this week
🔎 ML Research
MuJoCo
DeepMind published an insightful blog post describing the capabilities of MuJoCo, a physics simulator to advance robotics research →read more on DeepMind blog
Few-Shot Learning Training Dataset
Microsoft Research published a paper and open-sourced ORBIT dataset for few-shot learning models →read more on Microsoft Research blog
ML Underspecification
Google Research published a paper describing the risks of underspecification, a phenomenon that captures the mismatch between validation data and real-world behavior of ML models →read more on Google Research blog
🛠 Real World ML
Data Protection at Airbnb
The Airbnb engineering team discusses the architecture used to protect datasets across different data pipelines →read more on Airbnb blog
🤖 Cool AI Tech Releases
Using ML to Detect Code Bugs
Intel open-sourced Control Flag, a library that uses ML to detect bugs in code →read more on their GitHub page
A Toolkit for Robotics
Nvidia announced that Isaac, its framework for AI-powered robotics, has added support for the famous Robot Operating System (ROS) →read more in this coverage from The Robot Report
Data Shuffling in Spark
LinkedIn’s incubated Project Magnet, a framework for data shuffling framework that is now available in Apache Spark →read more on LinkedIn blog
💸 Money in AI
Computer memory and computer data storage producer Micron Technology unveiled its plans to invest $150 billion in leading-edge memory manufacturing over the coming decade.
To stay ahead in tech, NATO plans to adopt its first AI strategy and launch a $1 billion innovation fund to “futureproof” the 30-nation security pact. The AI strategy will integrate areas such as data analysis, imagery, and cyberdefense.
For ML&AI:
Mobile data intelligence startup Embrace raised $45 million in Series B funding led by New Enterprise Associates. Hiring in the US, Canada, Argentina.
Data management platform Metrolink.ai raised $22 million in a seed round led by Grove Ventures.
Deep learning development platform Deci raised $21 million in a Series A round led by Insight Partners. Hiring in Tel Aviv, Israel.
Collaborative data workspace Hex raised a $16 million Series A today led by Redpoint Ventures. Hiring in ‘US Timezones’.
Governance platform for AI Credo AI raised $5.5 million in a funding round led by Decibel, Village Global, and AI Fund.
Data insights and AI vendor Lyngo Analytics was acquired by enterprise data intelligence provider Alation Inc for an undisclosed sum.
AI-powered:
Maritime intelligence platform Saildrone raised a $100 million Series C round led by BOND. Hiring in Alameda, CA, US.
Automated document processing platform Rossum raised $100 million in a Series A round of funding led by General Catalyst. Hiring in Prague, Czech Republic.
Cybersecurity startup Cybereason raised a $50 million investment from Google Cloud. Hiring globally.
Workforce planning software SkyHive Technologies raised a $40 million Series B round led by holding company Eldridge. Hiring in Canada and the US.
Video intelligence platform Deep North raised $16.7 million in a Series A-1 round led by Celesta Capital and Yobi Partners.
Financial fraud detection platform Resistant AI raised $16.6 million in a Series A funding round from GV (formerly Google Ventures).
Threat detection marketplace SOC Prime raised a $11.5 million funding round led by DNX Ventures. Hiring in the US and Europe.
AI tool for product developers Mage raised $6.3 million in seed funding led by Gradient Ventures.
Conversational commerce platform Jumper.ai was acquired by business cloud communications provider Vonage for an undisclosed sum.