Artificial Intelligence

Using an AI Chess Program to Evaluate the Grand Masters

A new publication in ICGA Journal from Jean-Marc Alliot reports on using the AI chess program STOCKFISH to evaluate moves from 26,000 grandmaster games. In all, over 2 million positions were studied, with an aim at identifying the likelihood of a grandmaster making an incorrect move, as well as the magnitude of the error. Using a linear analysis, the results can be applied to rank all-time great grandmasters. At the top of the list is a group of four players: Magnus Carlsen, Vladimir Kramnik, Bobby Fischer, and Garry Kasparov. (April 21, 2017)

read more

Deep Learning for Classifying TB via Chest Radiography

Researchers from Thomas Jefferson University in Philadelphia have explored the use of convolutional deep learning neural networks for classifying tuberculosis from chest images. Performance was 96% accurate using a combination of methods and 99% accurate when combined...

read more

David Fogel Interviewed for Risk Roundup on Topic of Computational Intelligence

Dr. David Fogel was interviewed by Dr. Jayshree Pandya, founder of the Risk Group and host of Risk Roundup, a cyber-security & strategic security webcast/podcast. David spoke about various avenues for pursuing computational intelligence and the implications for future advancement in improving business operations, clinical research, and predictive analytics, among other topics. (April 23, 2017)

read more

New Doctoral Dissertation Studies the Ability to Detect Antisocial Behavior through NLP and Emotional Analysis

Myriam Munezero has published online the dissertation “Leveraging Emotion and Word-based Features for Antisocial Behavior Detection in User-Generated Content,” (Joensuu: University of Eastern Finland, 2017). The dissertation indicates that natural language processing methods combined with sentiment analysis can be useful in detecting anti-social behavior (ASB). Among the conclusions “It has been shown that ASBT contain a higher amount of swear words, profanity, insults, and violence-related words. ASBT contain both negative and positive emotions, though they contain more negative emotions, including anger. In addition, the high use of 2nd-person singular pronouns indicates that the target of ASB is someone other than the author. Further analysis could reveal the identity of targets. An improved understanding of the characteristics of ASB will also lead to improved intervention measures.”

read more

EffectCheck(R) Case Study of Oscar Munoz’ Letter to United Team

On April 11, 2017, United Airlines CEO Oscar Munoz offered an apology letter to the United “Team” (posted publicly on Facebook). Using EffectCheck, I analyzed the emotions evoked in the letter and have posted a case study of suggested improvements to the letter. The case study is also found at: https://www.linkedin.com/pulse/effectcheck-case-study-oscar-munoz-letter-united-team-david-fogel (April 19, 2017)

read more