InfoQ Homepage AI, ML & Data Engineering Content on InfoQ
-
OpenAI Announces GPT-3 AI Language Model with 175 Billion Parameters
A team of researchers from OpenAI recently published a paper describing GPT-3, a deep-learning model for natural-language with 175 billion parameters, 100x more than the previous version, GPT-2. The model is pre-trained on nearly half a trillion words and achieves state-of-the-art performance on several NLP benchmarks without fine-tuning.
-
QCon San Francisco 2020 Announces Program Committee
The QCon team has finalized the Program Committee for QCon San Francisco 2020 (Nov 16-18). The committee works on all aspects of software development. At QCon topics, track hosts and speakers are handpicked to guarantee relevant and timely content.
-
AWS Releases its Machine Learning Powered Enterprise Search Service Kendra into General Availability
Recently Amazon announced the general availability of its enterprise search service Kendra on AWS. With the GA release of Amazon Kendra, the public cloud provider added a few new specialized features and improved service accuracy.
-
Microsoft Build 2020: Highlights
Last week Microsoft held the 10th edition of Build, its annual conference aimed at developers using Microsoft technologies. The online event included multiple important announcements and releases, such as the general availability of Blazor WebAssembly, updates on the upcoming .NET 5, Azure Static Web Apps, and new projects related to IoT and Artificial Intelligence.
-
Google Open-Sources AI for Using Tabular Data to Answer Natural Language Questions
Google open-sourced Table Parser (TAPAS), a deep-learning system that can answer natural-language questions from tabular data. TAPAS was trained on 6.2 million tables extracted from Wikipedia and matches or exceeds state-of-the-art performance on several benchmarks.
-
Jepsen Disputes MongoDB’s Data Consistency Claims
In an article, MongoDB claimed their database passed “the industry’s toughest data safety, correctness, and consistency Tests”. In response, Jepsen published an article stating that MongoDB 3.6.4 had in fact failed their tests; the newer MongoDB 4.2.6 has more problems including “retrocausal transactions” where a transaction reverses order so that a read can see the result of a future write.
-
Amazon Announces General Availability of UltraWarm for Its Elastic Search Service on AWS
Recently, Amazon announced the general availability of UltraWarm for its Elasticsearch Service on AWS. Ultrawarm is a low cost warm storage tier, and extension to the Elasticsearch Service - offering up to three petabytes of storage, at almost a 90% cost reduction over existing options.
-
Google AI Launches NLU-Powered Tool to Help Explore COVID-19 Literature
Google AI launched COVID-19 Research Explorer, which provides a semantic search interface on top of the COVID-19 Open Research Dataset to help scientists and researchers efficiently analyze all of the dataset’s journal articles and preprints.
-
Mozilla Launches Hubs Cloud
Mozilla’s Mixed Reality group launches a cloud version of Mozilla Hubs, their social space for virtual reality gatherings. Organisations can now deploy and customize their own instance of Mozilla Hubs. It is available on the AWS Marketplace and manages all necessary AWS resources. Mozilla Hubs allows people to meet in a 3D environment, or join using a computer or a virtual reality device.
-
Google Open-Sources New Higher Performance TensorFlow Runtime
Google open-sourced the TensorFlow Runtime (TFRT), a new abstraction layer for their TensorFlow deep-learning framework that allows models to achieve better inference performance across different hardware platforms. Compared to the previous runtime, TFRT improves average inference latency by 28%.
-
Predicting Failing Tests with Machine Learning
Machine learning can be used to predict how tests behave on changes in the code. These predictions reduce the feedback time to developers by providing information at check-in time. Marco Achtziger & Dr. Gregor Endler presented how they are using machine learning to learn from failures at OOP 2020.
-
DeepMind's Agent57 Outperforms Humans on All Atari 2600 Games
Researchers at Google's DeepMind have produced a reinforcement-learning (RL) system called Agent57 that has scored above the human benchmark on all 57 Atari 2600 games in the Arcade Learning Environment. Agent57 is the first system to outperform humans on even the hardest games in the suite.
-
Amazon Introduces the New Streaming ETL Feature on AWS Glue
Recently, Amazon announced AWS Glue now supports streaming ETL. With this new feature, customers can easily set up continuous ingestion pipelines that prepare streaming data on the fly and make it available for analysis in seconds.
-
Google Updates Cloud Spanner with New Features: Backup on Demand, Local Emulator, and More
In a recent blog post, Google announced a slew of new features for Cloud Spanner. The public cloud provider released the service with new updates including the managed backup-restore feature, foreign key support, query optimizer versioning feature, a C++ client library, and a local emulator.
-
Blender, Facebook State-of-the-Art Human-Like Chatbot, Now Open Source
Blender is an open-domain chatbot developed at Facebook AI Research (FAIR), Facebook’s AI and machine learning division. According to FAIR, it is the first chatbot that has learned to blend several conversation skills, including the ability to show empathy and discuss nearly any topic, beating Google's chatbot in tests with human evaluators.