Tooploox CS and AI news 13

Tooploox CS and AI News #15

While not November news per se, the upcoming NeurIPS conference is arguably the hottest AI-related topic at the moment, with everything ready to start in the upcoming week. But there were several other topics to see and follow. 

November has brought multiple other interesting topics, with a new version of GauGAN and an AI-first drug discovery company being great examples of how the world is changing rapidly. 

Nvidia translates text to images with GauGAN2

Nvidia has delivered the GauGAN2, a tool that transforms various inputs, be they a sketch or text, into a lifelike image. The tool aims to deliver a photorealistic image from a mix of simple drawings and sentences or single descriptive words. This is the second generation of the tool, with the previous one being simply named GauGAN.

The name of the tool was inspired by Paul Gauguin, a French post-impressionist painter, whose images are far from photorealistic. Yet the painter was renowned for his experiments in the use of color. 

The tool can be used to modify images as well as bring them into existence from text or a vision alone. Nvidia claims that it will find multiple uses in supporting creative processes in filmmaking and graphic design, especially as a tool to be used in digital brainstorming sessions. The company claims that even the less polished, previous version of the tool was used in a such way. 

The tool’s demo can be found on a dedicated webpage. More information on the tool can be found on the Nvidia webpage.

AI startup funding hits record $17.9 billion in Q32021

Artificial Intelligence is currently one of the hottest topics on the planet – especially when it comes to the startup ecosystem. According to the State of AI report from CB Insights, global funding in the sector has seen an immense boost. The numbers have risen from $16.6 billion in 2Q2021 to $17.9 billion in just the third quarter. 

In Q3 alone there were 13 new AI unicorns with three startups reaching $2 billion in valuation. 

The full report is available on the CB insights webpage

Alphabet launches drug discovery company based on Deepmind’s protein folding research

The challenge of protein folding was one of those deemed unsolvable using traditional methods. There was no visible connection between the protein’s structure and the way it folds. On the other hand, though, the way that proteins fold has a tremendous impact on how it works in the organism. 

This challenge appeared to have been solved by AlphaFold2 – a model able to predict the way a particular protein will fold from its structure. To not lose momentum, Alphabet, the company behind Google and Deepmind (among others), decided to launch Isomorphic Labs. 

Leveraging the experience gathered during the AlphaFold project, the company aims to reinvent the drug discovery process by making it AI-first. By that, the company aims to uncover the “fundamental mechanisms of life.” 

The whole announcement can be found on the Isomorphic Labs website

Report: 95% of execs consider hiring AI talent a challenge

The execs agree that AI can bring significant advantages for their companies and building these solutions in an ethical way can add additional curves on the matter. 

On the other hand, though, the execs can see challenges in adopting AI, with the need to deliver proper training to their teams to get the full benefits of using AI-based solutions in their organizations. The need to provide training to their teams is seen as a challenge for companies.

The full version of the report can be found on the RELX website

Upcoming NeurIPS conference

Last but not least – there will be a NeurIPS conference in December. While it is not news about what happened in November, NeurIPS is one of the most important events in the AI world. 

The Conference on Neural Information Processing Systems (NeurIPS) has been running since 1986 when biologists and IT specialists would meet to exchange ideas and experiences regarding the building of neural network-inspired systems. With time, the conference has shifted from an interdisciplinary scientific meeting to more machine learning-focused events. 

Tooploox will also be present at the conference, having published two NeurIPS 2021 research papers. This year’s conference is not the first one Tooploox will support.

Similar Posts

See all posts