This newsletter is a biweekly curation of the wittiest trends from the world of Crypto, Blockchain, and AI. Subscribe here to get it directly in your inbox every Mondays and Fridays.
Here’s what’s happening…
Machine learning and AI can now create easily degradable plastics
- Plastic pollution accounts for a significant portion of environmental issues.
- On Thursday, October 21, University of Chicago researchers published a study, which shows a way through which modelling and machine learning can be combined to design plastics, and other materials, that can quickly degrade, without harming the environment.
- Apart from drastically reducing plastic pollution, their new development can allow companies and engineers to create materials with the specific properties they desire, with a high precision, in a more affordable and sustainable way.
LandingLens is the new platform that helps manufacturers visually inspect goods for defects
- Companies depend on humans to inspect goods for defects after manufacturing.
- Landing AI has launched a visual inspection platform called LandingLens
- The new platform will help product manufacturers build and deploy systems for visual inspection. Through this, manufacturers can now use AI to detect scratches on the screens of smartphones, defects in raw steel before it is used in production, etc.
- The platform is not limited to a particular kind of product, and it allows inspection experts from a manufacturing company to train a model on the input data of what a defect looks like. With this new AI-driven platform, any manufacturer can immediately know if their good is in good shape or not, without human supervision
AI beats humans at speech recognition for the first time
- In previous years, it has proven impossible for a computer system to transcribe a voice conversation as accurately as a human.
- For the first time ever, Karlsruhe Institute of Technology (KIT) researchers have successfully developed a computer system that performs better than humans in recognizing spontaneously spoken language in near real-time.
- With KIT’s new development, it will be possible to create computer systems that that act as live translators, translating from one language to another, with minimal latency, functioning far beyond human-level performance compared to the existing systems.
Lost languages can now be deciphered using machine learning
- Most languages that have ever been in existence are no longer spoken. A lot of these languages are also considered to be undeciphered or lost because we cannot understand writings of them.
- Researchers at the Computer Science and Artificial Intelligence Laboratory (CSAIL) in MIT have developed a new system that can decipher lost languages.
- Their system can uncover relationships between languages and identify language families.
- Using just a few thousand words, the system will be able to decipher languages that the world has lost for decades or possibly centuries.
Did you enjoy our curation? Don’t forget to subscribe to Decentralize Brief newsletter.
Have a lovely weekend!