Four short links: 13 January 2020
Simulated Customer, Symbolic Meets Statistical, Deep Fakes, and Online Radicalization
- Simulated Customer — The site will randomly generate one of 40 different [sales] objections, and give you 20 seconds to answer it.
- From Shallow to Deep Interactions Between Knowledge Representation, Reasoning, and Machine Learning — This paper proposes a tentative and original survey of meeting points between knowledge representation and reasoning (KRR) and machine learning (ML), two areas which have been developing quite separately in the last three decades. […] This paper is the first step of a work in progress aiming at a better mutual understanding of research in KRR and ML, and how they could cooperate.
- NHK Raises the Dead to Mixed Reviews — Enka singer Hibari Misora graced the “Kohaku” stage for the first time in decades to perform a new song. Well, technically, it wasn’t Misora herself—she died in 1989. Rather, it was a life-like hologram performing this fresh tune thanks to Yamaha’s Vocaloid: AI, a piece of technology that can replicate voices. Deepfaked audio and imagery. (via Hacker News)
- Empirical Studies of Online Radicalization: A Review and Discussion — Only 18 studies that met Desmarais et al.’s (2017) stringent systematic review criteria empirically examined the radicalization process. Fewer still, presumably, examined the online radicalization process. Indeed, Hassan et al., (2018) conducted a systematic review specifically focused on the relationship between the impact of extremist online content and violent radicalization. Eleven studies fit their eligibility criteria. […] The emerging evidence base is also pretty clear. Those who are radicalized and/or commit acts of terrorism have generally been exposed to radicalizing content. Exposure to this content leads to affective, emotional, and behavioral change at each stage of the process. Of course, some of these studies have relatively small sample sizes, and are only focused on specific types of terrorists or geographical contexts. The key now is to replicate and build upon this preliminary evidence to give us a sense of not just whether exposure to ideological content in the online environment causes violent extremism, but also how, in what contexts and for whom? Is “exposure” sufficient whether it is in the virtual or physical world? Does it work differently for different people in different contexts?