What’s Happening in Tech

Data Points

Vol. 15 // 2020


Microsoft has built an image-captioning algorithm that surpasses human accuracy in specific tests. The company utilized the AI system to update its assistant application for the visually impaired with plans to integrate it into Outlook, Word, Powerpoint, and other Microsoft products. According to researchers, the model uses VIVO that utilizes paired image-tag data to learn visual vocabulary. To benchmark performance, Microsoft entered the new AI system into the ‘nocaps’ challenge and currently ranks first on its leaderboard. (“Microsoft’s new image-captioning AI will help accessibility in Word, Outlook, and beyond” – The Verge)


The AI-powered GPT-3 technology is engineered for content marketers looking to score well on Google with high-ranking blog posts. Users only need to enter keywords into the search bar and the system will evaluate top-performing content associated with the subject on Google. It then generates a unique idea by running the data through the GPT-3 text generator. GPT-3 can work on anything that integrates a language structure. This includes writing essays, answering queries, summarizing lengthy texts, translating languages, and even creating computer code. (“What Is GPT-3 And Why Is It Revolutionizing Artificial Intelligence?” – Forbes)


Researchers from IBM and Pfizer have come up with an innovative AI model that predicts Alzheimer’s at an early stage in healthy individuals better than basic clinical tests. It leverages a short language sample from a verbal cognitive test that a patient takes. The AI model can predict the onset of Alzheimer’s with 71% percent accuracy, compared to standard clinical tests that are correct roughly 59% of the time. This study can result in the creation of an easily accessible metric that healthcare professionals can use to evaluate the risk of Alzheimer’s, making timely intervention possible. (“5 Things to Know About IBM’s AI Predictor for Alzheimer’s Disease” – IBM News Room)


NVIDIA has another contribution to its name as its AI inference in MLPerf with A100 Tensor Core GPUs set records on data center, cloud, and edge benchmarks. The latest benchmarks presented four new tests that gauge performance in NLP, recommendation systems, speech recognition, medical imaging, and computer vision AI use cases. (“NVIDIA Breaks 16 AI Performance Records in Latest MLPerf Benchmarks” – NVIDIA)


Intelligent automation specialist Eggplant announced the launch of its new AI-powered software testing platform. It is designed to enhance the quality of digital experiences and the speed of deployment on any platform. Enterprises benefit from improved levels of automation and faster time to market that quickens the pace of the continuous delivery approach. Organizations can seamlessly deploy this solution on Eggplant’s DAI platform that features cloud-based automation, monitoring insights, and fully-automated, self-healing test assets. (“Eggplant Launches Cloud Platform for AI-Driven Intelligent Test Automation” – businesswire)

Related Articles

Turing Talk