By: Sterling Payne, B.Sc.
To Beat Go Champion, Google’s Program Needed a Human Army
“It may be a hundred years before a computer beats humans at Go — maybe even longer,” Dr. Piet Hut communicated to New York Times’ George Johnson in a 1997 conversation. The event prompting their discussion was the victory of IBM’s Deep Blue over grandmaster Garry Kasparov in a series of chess games. Dr. Hut’s prediction was bested by about 80 years by AlphaGo, the product of Google’s DeepMind. AlphaGo recently secured a victory against 9 dan Go champion Lee Sedol in a 5-game match hosted by Google. By nature, Go as a game is more complex than chess; less stringent gameplay guidelines don’t offer a surefire way to determine which player is at an advantage. Rather than powering through an analysis of thousands upon thousands of potential moves each turn, AlphaGo utilizes a novel combination of machine-learning methods to determine which board configurations are more advantageous, and positively reinforces correct decisions via thousands of matches played against itself. The product of this is an artificial intelligence (AI) that more closely represents human intuition, at least in the small scope of the Chinese board game.
With its 4-1 victory over Sedol, AlphaGo demonstrated extreme proficiency in the game of Go, but in only that. While inarguably an astounding accomplishment and significant leap in the field of computer science, AIs like AlphaGo have a long way to go before they can replicate the intuition of the human mind, which is far expandable beyond an ancient board game. In terms of policy, the very methods used to create AlphaGo could also find their ways into hospitals and healthcare facilities in the near future. With the advent of artificial intelligence in the workplace, extra considerations will have to be taken by patients and care providers alike in terms of personal information, data management, and general communication. (George Johnson, The New York Times) (Will Knight, MIT Technology Review)
Federal Cancer Research
Blue Ribbon Panel Announced to Help Guide Vice President Biden’s National Cancer Moonshot
The Cancer Moonshot Initiative , headed by Vice President Joe Biden, plans to put an end to the disease that has plagued millions of humans for hundreds of years. Armed with a $1 billion budget over the next five years, the initiative’s primary aim is to speed up cancer research such that a decade’s worth of discoveries can occur in half that time. Two of the main areas where such discoveries will fall are detection and treatment. A task force to handle financial matters and progression of the initiative was announced in February, and just yesterday (April 04, 2016), the National Cancer Institute unveiled their Blue Ribbon Panel, a special selection of various leaders in the fields of cancer research and patient advocacy, to direct efforts of the initiative to where they are likely to make the largest impact.
As a society, our knowledge of cancer has grown considerably since the turn of the century; Cancer is no longer thought of as a single disease that affects people, rather, it is the product of multiple genetic mutations and cellular microenvironments, painting a unique disease landscape for each person it affects. Members were chosen such that the panel represented multiple walks of science from immunology to bioinformatics, as well as cancer prevention and treatment. Already armed with capital and a team to guide finances and general progress, the Cancer Moonshot Initiative has taken another giant step forward with the addition of the Blue Ribbon Panel. The full member list of the Blue Ribbon Panel and the original announcement are linked here. (News Releases, National Institutes of Health)
Biology software promises easier way to program living cells
With computer programming, the programmer gives the computer a set of instructions in one (or more) of several different programming languages. These instructions include logical operations such as true-false statements (i.e. “if this is true, then do this”) and various loops (i.e. “while this is true, do this”). At the end of all of this, sits a program, executed by the computer to provide some sort of output, whether it be ordering a data set, turning on a light, or spinning a motor. Dr. Christopher Voigt and his lab at MIT have taken these principles and applied them to their new software Cello, a programming language capable of producing working circuits in living systems. Cello requires the user to input commands, such as a function they would like a given cell to perform and under what conditions it should perform said function. After the input is compiled, the end result is a DNA sequence or “circuit” that, when placed inside a cell, can fulfill the function(s) specified by the user. In a paper recently published in Science (April 01, 2016), Alec Nielsen and colleagues used cell to generate 60 different DNA circuits, 75% of which worked as expected the first time when introduced into Escherichia coli cells.
As synthetic biology continues to grow and gain popularity throughout the research world, it is of increasing importance to think about what policies and potential restrictions should be set in place. Engineering de novo biological systems and functions can be extremely powerful, yet, if left in the wrong hands, could have significant consequences as with any equally commensurate technique (e.g. CRISPR-Cas9). (Erika Check Hayden, Nature News)
Have an interesting science policy link? Share it in the comments!