ℹ️🅿️🌀 The First AI Startups IPOs
Weekly newsletter that discusses impactful ML research papers, cool tech releases, the money in AI, and real-life implementations
📝 Editorial
Initial public offerings (IPOs) are not only a major achievement in the lifetime of any company, but also a strong sign of maturity in a given market. In recent technology trends, the first group of startups that make it to the IPO line have gone to become emblematic companies of those respective markets, while also signaling that the market has matured enough to produce standalone viable companies. The examples are everywhere: Salesforce.com in cloud, Okta in enterprise identity, Cloudera in big data, New Relic in application performance monitoring, Twilio in cloud communications; the list goes on and on. The first IPOs of any tech trend are not necessarily the most successful companies, but they certainly pave the wave for the rest of the market. In artificial intelligence (AI), several public companies such as Nvidia, Microsoft, and Alphabet have capitalized on the trend, but we still haven’t seen startups from the AI-era debut as public companies.
That’s about to change.
The rapid growth of the AI space is accelerating the path of several high flying startups to become publicly traded companies. Just this week, Tom Siebel’s C3.ai filed IPO prospectus under the ticker symbol “AI” (lucky them). Also, AutoML platform startup DataRobot raised a monster $270 million round that is signaling its intention to go public in the near future. This first wave of IPOs will test the market sentiment with respect to AI trends and open the door to a new generation of AI-first publicly traded companies.
What do you think? Are public markets ready for new AI startups?
🔺🔻TheSequence Scope – our Sunday edition with the industry’s development overview – is free. To receive high-quality educational content every Tuesday and Thursday, please subscribe to TheSequence Edge 🔺🔻
🗓 Next week in TheSequence Edge:
Edge#41: the concept of Long-Short Term Memory Networks (LSTMs); one of the biggest breakthroughs in AI history by OpenAI; Uber Manifold.
Edge#42: the review of LinkedIn’s Dagli, a new ML Framework for Java Developers.
Now, let’s review the most important developments in the AI industry this week.
🔎 ML Research
Understanding Uncertainty
Berkeley AI Research Lab (BAIR) published a research paper explaining the concept of Amortized Conditional Normalized Maximum Likelihood, a new metric to quantify uncertainty in classification models ->read more on BAIR blog
MinDiff
Google Research published a paper outlining MinDiff, a new method for mitigating unfair bias during machine learning training ->read more on Google Research blog
Facebook and Harmful Content
Facebook AI Research (FAIR) published a series of blog posts about the machine learning methods used to handle different forms of harmful content ->read more in the FAIR blog posts about detecting hate speech and misinformation
🤖 Cool AI Tech Releases
DeepMind Lab2D
DeepMind open-sourced Lab2D, a system for advancing research in multi-agent reinforcement learning systems ->read more on their GitHub page
Mac Optimized TensorFlow
Apple released a version of the TensorFlow deep learning framework optimized for Mac hardware ->read more on TensorFlow blog
The Language Interpretability Tool (LIT)
Google open-sourced the LIT, an interactive toolkit that addresses challenges specific to NLP models, helping explore and analyze their behavior ->read more on Google AI blog
💬 Useful Tweet
New York University Center for Data Science (NYU-DS) released the materials used in AI legend Yann LeCun’s deep learning course
💸 Money in AI
Platforms that help automate the building, deploying, and managing of machine learning models are in the spotlight this week.
AutoML platform DataRobot raised $270 million in an equity funding round. The platform democratizes data science for enterprises, providing them with end-to-end automation for building, deploying, and managing ML models.
Deep learning models automation startup Abacus.AI raised $22 million. The company leverages technologies such as GANs (Edge#8) and NAS (Edge#4) to simplify building AI models in order to create large-scale, real-time customizable deep learning systems.
BeyondMinds announced a $15 million round. The startup offers a modular AI technology stack to facilitate enterprise product deployments.
Cloud-agnostic Seldon, another ML deployment startup, raised $9.4 million.
MLOps platform Arrikto raised $10 million. The startup tries to speed up the ML development lifecycle by allowing engineers and data scientists to treat data like code.
Software and services company Autodesk acquired cloud-based AI software for urban development startup Spacemaker for $240 million. Autodesk said that “Spacemaker is in line with the company’s long-term strategy of using the power of the cloud, ‘cheap computing' and machine learning to evolve and change the way people design things.”
AI-powered fraud detection platform for e-commerce Forter has raised $125 million in a Series E round. Using ML and behavioral analytics, Forter builds customers’ portraits and their intent, flagging transactions if they look suspicious.
Investors are also seeing lots of potential in telehealth and biometrics:
AI-based personal ECG technology and provider of enterprise cardiology solutions AliveCor raised $65 million in funding. Their algorithms detect atrial fibrillation, bradycardia, tachycardia, and other health issues from heart rate readings.
AI telemedicine app K Health closed a $42 million Series D round. The startup leverages AI to source a massive database of anonymized reports in order to diagnose health issues.
AI-powered glycoproteomics startup InterVenn Biosciences raised $34 million. The company claims its AI-imbued product automates the discovery of biomarkers (indicators of the severity of some diseases) and even the design of certain clinical trials.
Health care data science startup ClosedLoop.ai raised $11 million in Series A funding. The platform supports data science teams with healthcare-specific tools for data prep, automated feature engineering, AutoML / model training, and deployment / MLOps.
Quantum software startup Zapata Computing raised $38 million. One of the nearest-term quantum use cases will be in machine learning. Zapata’s recently launched hardware-agnostic quantum computing platform Orquestra allows organizations to leverage quantum capabilities to generate augmented data sets, speed up data analysis, and construct better data models for a range of applications. Quantum hardware can, in theory, reduce the training time of deep networks from months to hours.
Computer vision AI platform Chooch Al closed a $20 million Series A round. Its platform replicates human visual tasks and processes using a complete computer vision deployment process across a wide variety of industries.
AI-driven call center platform Cogito raised $25 million. Using natural language processing (NLP), the platform analyzes conversations measuring energy level, pace, tone of speech, and other factors to capture and interpret speakers’ intent, helping them recognize mistakes and make corrections on the fly.
AI-driven safety system for motorcyclists Ride Vision stealthily emerged with a $7 million round. Ride Vision uses a combination of image-recognition and AI technologies to power its predictive vision algorithms that help riders make critical life-saving decisions in real-time.
Want to contribute? Forward TheSequence to a friend or share this post
TheSequence is a summary of groundbreaking ML research papers, engaging explanations of ML concepts, exploration of new ML frameworks, and platforms. It also keeps you up to date with the news, trends, and technology developments in the AI field.
5 minutes of your time, 3 times a week– you will steadily become knowledgeable about everything happening in the AI space.
Keeping an eye on Butterfly Network to see how the markets react to hardware+cloud+AI.