Researchers propose paradigm that trains AI agents through evolution

#artificialintelligence

A paper published by researchers at Carnegie Mellon University, San Francisco research firm OpenAI, Facebook AI Research, the University of California at Berkeley, and Shanghai Jiao Tong University describes a paradigm that scales up multi-agent reinforcement learning, where AI models learn by having agents interact within an environment such that the agent population increases in size over time. By maintaining sets of agents in each training stage and performing mix-and-match and fine-tuning steps over these sets, the coauthors say the paradigm -- Evolutionary Population Curriculum -- is able to promote agents with the best adaptability to the next stage. In computer science, evolutionary computation is the family of algorithms for global optimization inspired by biological evolution. Instead of following explicit mathematical gradients, these models generate variants, test them, and retain the top performers. They've shown promise in early work by OpenAI, Google, Uber, and others, but they're somewhat tough to prototype because there's a dearth of tools targeting evolutionary algorithms and natural evolution strategies (NES).


How Artificial Intelligence is transforming eCommerce marketers in 2020

#artificialintelligence

Artificial Intelligence has grown massively over the past few years. You would have noticed how it is making the businesses strong and sufficient with its capabilities and power. From managing the customers at your store to improving their overall experience to providing them with personalized solutions, there is a lot that the technology can do for your eCommerce outlet. AI, though it is implemented in most businesses, is still a futuristic technology. According to a report by PwC, this technology will contribute close to $15.7Tn by 2030 to the global economy.


Addressing Drawbacks Of AutoML With AutoML-Zero

#artificialintelligence

Automated machine learning – or AutoML – is an approach that cuts down the time spent in doing iterative tasks concerning model development. AutoML tools help developers build scalable models with great ease and minimal domain expertise. AutoML is one of the most actively researched spaces in the ML community. AutoML studies have discovered ways to constrain search spaces to isolated algorithmic aspects. This includes the learning rule used during backpropagation, the gating structure of an LSTM, or the data augmentation.


Machine Learning Tokyo

#artificialintelligence

Episode 80: The featured guests are Machine Learning Tokyo (MLT) members Suzana Ilić (co-founder), Dimitris Katsios, and Asir Saeed. MLT is a Tokyo-based nonprofit organization dedicated to democratizing Machine Learning (ML). They are a team of engineers and researchers--now with a community of 5,000 people--and the winner of the Rakuten Technology & Innovation Silver Award 2019.





The future of contact tracing GovInsider

#artificialintelligence

When an epidemic spreads, it's vital to alert people who may have been infected but are yet to show symptoms. This process is called contact tracing, and across the world officials are spending long nights and weekends interviewing patients to identify people at risk. "The challenge is that, right now, it takes a significant amount of time to identify close contacts of the person who has been infected," says Akshay Saigal, Head of Innovation Labs for Asia at DXC Technology. "The track and trace is very manual and takes huge teams of dedicated officials to perform." He thinks that tech innovation can make a big difference to nations in need and spoke with GovInsider about three steps that officials can use for faster contact tracing.



Could AI Security Prevent Hacks?

#artificialintelligence

Las Vegas hacking event, the Cyber Grand Challenge was the ultimate, and only, all-machine hacking competition. Each machine identified software vulnerabilities, exploited them and patched their own systems to protect against threats -- all without a human programmer intervening. This article explains the role of automation in IT security and how it could address the skills shortage. We've all heard of the wider IT skills shortage, but the lack of security skills in the industry is even more critical. According to a report by the Life and Times of Cybersecurity Professionals, IT workers that have specialist cyber security skills are approached with a new job offer at least once a week.