The advent of AlphaGo AI, which managed to beat the world’s best human Go player, has render Google-owned corporation DeepMind, its inventor, renowned and respected. Several years after AlphaGo’s success, DeepMind now puts forward AlphaFold, a complex system aiming to predict protein structure for biologists and experimentalists.
Anticipating a biological structure has never been an ready task, especially for proteins containing an abundance of amino acids and chemical bonds peculiar to each of them. AlphaFold, in this case, allows scientists to accurately predict the 3-D structure of up to 200 million proteins with artificial intelligence and machine learning, significantly extending the boundaries for biology and drug development.
Compared to pure engineering process, AlphaFold distinguishes itself from the field by combining engineering and machine learning technology with various disciplines in biology and chemistry. Basically, 32 algorithms together constitute AlphaFold and it is able to estimate the angles between chemical bonds through learning from precedent plausible predictions.
It is hard to imagine that about 10 years ago, the term of AI literally has little notice, yet nowadays every individual may have some words to discuss about it. And the society has made huge strides thank to the efficient adoption of AI. AlphaFold has been diversely applied to fields of antibiotic manufacture, plastic-eating enzymes, augmented crops, etc. It is also widely playing a role in pharmacy, which contributes to the generation of better cures for the society.
However, ethical issues still arise and they stem from the nature that AlphaFold is incredibly powerful and potentially toxic. Scientists in DeepMind bear the phrase of “pioneering responsibly” in mind and agree on not fully releasing the technology until developers manage to thoughtfully check the application of AlphaFold models.
Just as the co-founder of DeepMind Demis Hassabis expressed in an interview, “it’s a dual-use technology—it depends on how, as a society, we decide to deploy it—and what we use it for.”
Resources:
https://www.scientificamerican.com/article/one-of-the-biggest-problems-in-biology-has-finally-been-solved/
unsplash.com