The Association for Computing Machinery (ACM) announced the latest A. M. Turing Award recipients at the end of March, to honor their contributions towards the breakthrough of artificial intelligence (AI). The Award is frequently recognized as the “Nobel Prize of Computing”. It was named after Alan M. Turing, the English mathematician who developed influential algorithms and concepts which gave birth to modern computers.
The prestigious prize, a total of US$1-million, was shared between Yoshua Bengio, Geoffrey Hinton and Yann LeCun. ACM called them “the fathers of deep learning revolution”. The trio had been working individually and collaboratively, on various works that formed the foundations of deep neural networks. Neural Networks set itself apart from other computer models. The circuits of neurons, their non-linear connections
The Montreal declaration and AI concerns
Presently, Bengio is a professor at the University of Montreal and the scientific director of Montreal Institute for Learning Algorithms (MILA). Hinton works for Google, as the vice-president and an engineering fellow while LeCun is the vice-president and chief AI scientist at Facebook and a professor at New York University.
Last December, Bengio presented a set of guidelines to urge a responsible development of AI called the Montreal declaration. In an interview with Nature this January, Bengio expressed his concerns towards the misuse of AI and new technology. He feels that AI has aggravated inequality as those in power are using the technology to keep, or even strengthen that power. On the other hand, prejudices continue as people train AI using data that reflect their own thoughts and believes, which tend to be biased.
Bengio believes misuse will persist and most of them will take place secretly. Thus, there is a need to raise concerns before some of them happen and these become the force behind putting forward the initiative. Bengio said the Montreal declaration is unique as it does not include only scholars, interested enterprises and members of public can also read and sign the declaration online.
Surviving the “AI winter” and lessons for healthcare and medicine
Bengio, Hinton and LeCun’s successes were not overnight. They picked up the technology during “AI winter”. AI is often trapped in cycles of hype and bust. It was a dark period in the mid-90s and early 2000 when people lost interest in AI. There were not many sponsorships towards research relating to computing or computational sciences. The trio met in this winter, they began to exchange ideas on neural networks and how it contributes to character recognition. Their work did no catch attention until 2012.
If you have attended the latest AIMed breakfast briefing: Experience the future of AI in Radiology (Chicago), you probably had caught up with the energetic and practical presentation made by Dr. Paul J. Chang, professor
This is why no healthcare professionals will regard electronic health records as intuitive or an invention which matches with the present technology. Because by the time, AI in medicine catches with the present AI development, AI in other industries will have also make their respective advancements. As such, if AI surely moves in cycles of hype and bust, then the AI medicine community should be prepared for a possible “AI winter” in near future. This is especially so infrastructures in most healthcare institutions are still immature to support AI and professionals are skeptical if it will be fitted into their workflow.
In another interview with MIT Technology Review, said the most significant difference between now and in the 90s is the availability of data. This is a big step towards making AI more democratic. In order to ensure AI will continue to be used in the interests of human, there is a need to collaborate. AI development should not be a scientific race.
A science writer with data background and an interest in current affair, culture and arts; a no-med from an (almost) all-med family. Follow on Twitter.