The Sequence Scope: The Microsoft AI Powerhouse
Weekly newsletter that discusses impactful ML research papers, cool tech releases, the money in AI, and real-life implementations.
The Sequence Scope is a summary of the most important published research papers, released technology and startup news in the AI ecosystem in the last week. This compendium is part of TheSequence newsletter. Data scientists, scholars, and developers from Microsoft Research, Intel Corporation, Linux Foundation AI, Google, Lockheed Martin, Cardiff University, Mellon College of Science, Warsaw University of Technology, Universitat Politècnica de València and other companies and universities are already subscribed to TheSequence.
📝 Editorial: The Microsoft AI Powerhouse
Microsoft’s transformation in the last few years under Satya Nadella is one of the best stories of technological innovation in corporate America. And artificial intelligence (AI) has been at the center of it. AI has become one of the trends that catalyzed the adoption of Microsoft technologies. Microsoft’s impressive success in AI has different building blocks. Microsoft Research has been advancing AI research in very diverse areas. The Azure platform has been powering native services to enable AI capabilities boosting a strong developer and partner communities. Microsoft remains competitive in AI markets like digital assistants and gaming and has been incredibly smart about acquisitions in the AI market.
Earlier this year, Microsoft acquired robotic process automation (RPA) startup Softmotive to bolster its Power Automate Platform. The move has borne fruit: Today, Microsoft unveiled Power Automate Desktop, an app developed atop Softmotive’s RPA technology that provides desktop automation features for developers and non-coders alike.
During its Ignite conference this week, Microsoft unveiled a series of new capabilities to several of its AI services. The Microsoft Cognitive Services API launched the general availability of new services in areas such as anomaly detection or spatial analysis. Azure ML presented a brand new user interface to streamline the lifecycle of machine learning models. Other services such as the Azure Bot services also launched new capabilities. When you put together the reach and depth of AI solutions in the Azure platform with the amazing wave of AI open-source releases from its research unit, Microsoft starts looking like a front runner in the AI race.
Do you agree? How would you compare Microsoft against Amazon and Google in terms of AI offerings and capabilities?
Building on that trend, Microsoft today previewed improvements coming to Microsoft Search and Project Cortex — two enterprise search platforms powered by AI — and announced the public preview of Microsoft Power Automate Desktop. In addition to powering new Microsoft Teams and Windows experiences, Microsoft Search will expand its reach across first- and third-party services to include integration with Azure Cognitive Search and connections to content from over 100 external sources. As for Project Cortex, it’s splintering off into a product calledSharePoint Syntex — with several more to come.
Microsoft announced updates to Azure Cognitive Services and Azure Machine Learning aimed at streamlining business processes during the coronavirus pandemic. The company also launched two features in Azure Cognitive Search — Private Endpoints and Managed Identities — plus enhancements to Bot Framework Composer and the broader Azure Bot Service.
At Ignite 2020, Microsoft today announced a new communications offering called Azure Communication Services that’s built on top of Azure. Microsoft says it leverages the same network powering Microsoft Teams to enable developers to add multimodal messaging to apps and websites while tapping into services like Azure Cognitive Services for translation, sentiment analysis, and more.
Share
🔺🔻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#25: the concept of representation learning; an overview of Microsoft Research’s paper about representation and multi-task learning in language; Facebook’s fastText framework.
Edge#26: the concept of self-supervised learning; an overview of the self-supervised method for image classification proposed by Facebook; Google’s SimCLR framework for advancing contrastive self-supervised learning.
Now, let’s review the most important developments in the AI industry this week.
🔎 ML Research
Efficient NLP with Minimum Size
A paper by Google Research proposing a method that can achieve excellent performance in text classification tasks with a minimum model size ->read more on Google Research blog
A Dataflow Approach to Conversational AI
Microsoft Research presented a paper about a representation framework for modeling dialogues as graphs, demonstrating a new approch to conversational AI ->read more on Microsoft Research blog
The Magic Behind the New Alexa Features
The Amazon Research team has been publishing a series of insightful posts about the AI behind new Alexa’s capabilities ->read more in their posts about speaking style adaptation, natural turn talking and interactive teaching
🤖 Cool AI Tech Releases
DynaBench
Facebook AI Research (FAIR) unveiled Dynabench, a platform that takes a new approach to data collection and benchmarking for AI models ->read more on the FAIR team blog
Updates and releases from Microsoft
Microsoft announced updates to its Azure ML and Azure Cognitive Services platforms ->read more in this analysis from VentureBeat. Microsoft also released InnerEye Deep Learning Toolkit as open-source software, it allows to train train high-performance models for medical imagining https://github.com/microsoft/InnerEye-DeepLearning
KILT
Facebook AI Research (FAIR) open-sourced KILT, a new set of datasets and benchmark for knowledge-intensive language tasks ->read more on FAIR blog
GPT-3 for Microsoft
OpenAI licensed GPT-3 technology to Microsoft for their own products and services. GPT-3 still remains in limited data for researchers and partners ->read more on OpenAI blog
we are releasing the . The toolkit makes it easy to train high-performance models with ensembles and deploy your models using Azure Machine Learning or Azure Stack Hub. https://www.microsoft.com/en-us/research/blog/project-innereye-open-source-deep-learning-toolkit-democratizing-medical-imaging-ai/
💬 Useful Tweet
#DataScientist Job Market 2020 #KDN Data Scientist Job Market 2020 — KDnuggetsWith an analysis of over a thousand Data Scientist job descriptions in the USA, check out the trends for 2020 and current expectations on new positions in the field, including credentials, experience, and programming languages.ow.ly
September 26th 2020
3 Retweets3 Likes
💸 Money in AI
data intelligence company Ripjar, a U.K. company founded by five data scientists who previously worked together in British intelligence at the Government Communications Headquarters (GCHQ, the U.K.’s equivalent of the NSA), has raised $36.8 million (£28 million) in a Series B. Utilising proprietary artificial intelligence technology, Labyrinth uncovers new risks, and reduces false positives by over 90%. Labyrinth enables human analysts to make better informed decisions with speed and accuracy, saving significant cost and time for its clients.
machine learning and artificial intelligence. We believe that through this augmented analysis we can protect global companies and governments from the ever-present threat of money laundering, fraud, cyber-crime and terrorism.
- WhyLabs model and data monitoring AI startup WhyLabs came out of stealth mode with a platform for monitoring data and AI models >read more in their blog post AI model failures WhyLabs is launching out of stealth today with $4 million to grow its platform for data scientists who need help monitoring and troubleshooting problems they encounter with datasets or AI models.
- EasySend has just announced our $11M Round A Funding,No-code and low-code software have become increasingly popular ways for companies — especially those that don’t count technology as part of their DNA — to bring in more updated IT processes without the heavy lifting needed to build and integrate services from the ground up. EasySend, an Israeli startup which has built a no-code platform for insurance companies and other regulated businesses to build out forms and other interfaces to take in customer information and subsequently use AI systems to process it more efficiently, is announcing that it has raised $16 million.
- Connected car analytics startup Aurora Labs today raised $23 million has secured $23 million in Series B funding Aurora Labs has reinvented how automotive companies, Tier-1s, silicon vendors and enterprises conduct remote software management, remote diagnostics and over-the-air software updates. In addition to keeping software safe and secure from faults and cybersecurity attacks, Aurora Labs’ solutions allow manufacturers to continuously add new features and functions extending the life of the device and enhancing user experiences. All of this is accomplished with AI and machine learning technologies that will save manufacturers billions of dollars on their bill of materials and data communications costs and enable recurring revenue streams.
- AI-powered customer experience management platform Sprinklr has raised $200 million (kudos to our subscribers from Sprinklr 👏). Sprinklr’s “AI listening processing” solution allows companies to get structured and meaningful sentiments and insights from unstructured customer data that comes from public conversations on different websites and social platforms.
