The Year 2017 In Review
One of my leisure projects is the website "21st Century Headlines" where I post links to news items, reports and papers about biomedical sciences, computers and physics on an almost daily basis. Once a year I write a review. This is an abbreviated version of the review of the year 2017. The full review can be found online.
Written by Claus Volko
Vienna, Austria, Europe
Contact: cdvolko (at) gmail (dot) com
This is the second time I am writing an annual review for the website "21st Century Headlines - Things That Matter". Indeed, it is necessary that such a thing is done at least once a year, facing the sheer amount of science and technology news. It is important to occasionally stop just collecting and reading news in order to reflect on what has happened in the past year and evaluate what is really important.
Machine Learning and Artificial Intelligence
Statements of Political and Economic Leaders
This year, the media contained much more reports on machine learning and artificial intelligence than in the previous year, reflecting the public interest in this exciting branch of computer science. Many observers consider AI "the most important technology of our time", and even high-rank politicians such as Russian President Vladimir Putin ("What Putin Has To Say About AI?", "Putin says the nation that leads in AI 'will be the ruler of the world'") and the former US Presidential Candidate Hillary Clinton ("'Artificial intelligence is not our friend:' Hillary Clinton is worried about the future of technology"), entrepreneurs such as Elon Musk ("Elon Musk says artificial intelligence arms race most likely to cause World War III", "Podcast: Life 3.0 - Being Human in the Age of Artificial Intelligence", "If Humans Are to Survive, We Must Merge With Machines") and Bill Gates ("This Will Be the Biggest Technological Breakthrough of Our Lifetime") and also scientists such as Richard Dawkins ("A.I. Might Run the World Better Than Humans Do") expressed their views, expectations and concerns about it in public.
One of the latest milestones is AlphaGo Zero, developed by DeepMind, a subsidiary of Alphabet (the corporation which Google belongs to). While its predecessor AlphaGo made it to the press in 2015 for beating the human world champion in Go, AlphaGo Zero is a step ahead, as it "learns to learn entirely on its own" and "becomes Go champion without human input". Quote from the article "AlphaGo Zero: Learning from scratch": "Previous versions of AlphaGo initially trained on thousands of human amateur and professional games to learn how to play Go. AlphaGo Zero skips this step and learns to play simply by playing games against itself, starting from completely random play. In doing so, it quickly surpassed human level of play and defeated the previously published champion-defeating version of AlphaGo by 100 games to 0."
As EdgyLabs reported, "New AlphaGo Zero 'Unsupervised' AI is 100X Better While Using 10% Computing Power". MIT Technology Review even stated: "AlphaGo Zero Shows Machines Can Become Superhuman Without Any Help", and Carlos E. Perez, the author of Artificial Intuition and the Deep Learning Playbook, wrote on Medium that "AlphaGo Zero is a Quantum Leap Forward in Deep Learning".
In December a video documentary about AlphaGo was released in the Google Play Store.
Many articles have been published about possible benefits AI will bring to healthcare and other aspects of human life, such as the article "How artificial intelligence will save lives in the 21st century". More concretely, Insilico Medicine reported in July that the "first molecules discovered by next-generation artificial intelligence [are] to be developed into drugs" and that they are using artificial intelligence "to fight premature aging". That's exciting news, just like the promise of "beating cancer sooner with data and artificial intelligence".
There is also an article that boldly states: "Artificial Intelligence is the Stethoscope of the 21st Century" and another one expressing the view: "Artificial Intelligence Will Redesign Healthcare". The statements "There Is No Precision Medicine Without Artificial Intelligence" and "The rise of artificial intelligence means doctors must redefine what they do" (the latter written by Bryan Vartabedian, an assistant professor of pediatrics) are also to be understood in this context.
The website Medical Futurist also published an article about "Six Challenges To Tackle Before Artificial Intelligence Redesigns Healthcare". Also, MIT Technology Review published the article: "The Machines Are Getting Ready to Play Doctor", while techemergence wrote about "How America's 5 Top Hospitals are Using Machine Learning Today" as well as about "Machine Learning in Genomics - Current Efforts and Future Applications". Regarding genome research, there is also an article written by Francis Collins, the director of the National Institutes of Health, called "Creative Minds: Using Machine Learning to Understand Genome Function".
Artificial intelligence has also been successfully used in physics, as the article "The Neural Network: How artificial intelligence is fuelling 'Phasebook'" explains. Video game enthusiasts might be interested in the article: "Artificial intelligence uses less than two minutes of videogame footage to recreate game engine". Regarding applications in natural language processing, you might be interested in the articles "Artificial intelligence goes bilingual - without a dictionary" and "Natural Language Processing Projects & Startups to watch in 2017". Google is also trying to "reinvent chat with machine learning". In this context, Thomas Hornigold published the article "How Close Is Turing's Dream of Conversational Machines?" on the website Singularity Hub.
In order to "accelerate artificial intelligence", hardware manufacturers such as Intel are developing self-learning chips (Google is doing something similar) while researchers at the California NanoSystems Institute are trying to build computers that "match some of the brain's computational and energy efficiency with systems that mirror the brain's structure".
Not totally unrelated to this, you might also be interested in Tara Karimi's talk about the "Next Generation of Biologically Inspired Artificial Intelligence" as well as the articles "Machines Are Getting Smarter - Now They Should Explain Themselves" and "What Intelligent Machines Need to Learn From the Neocortex".
Riza C. Berkan also published an article about the question "Can Machine Learning Use Knowledge instead of Data?" in which he compares the methods of deep cloning and deep learning. Similar to this, Vincenzo Lomonaco wrote about continuous learning, which he deems "the key towards Machine Intelligence".
One of the future challenges is to get conscious machines, about which Singularity Hub featured an article.
But there are also some critical voices, such as the one of the former head of Uber's AI Labs, Gary Marcus, who "warns that the field isn't moving nearly as fast as many people think", or as the one of Ron Miller, who writes that "artificial intelligence is not as smart as you (or Elon Musk) think". Also, Oren Etzinio, the chief executive of the Allen Institue for Artificial Intelligence, pondered on the question whether society should regulate artificial intelligence and proposed "three rules for artificial intelligence systems". The articles by Shamli Prakash, "Man Vs. Machine: Understanding the Perils of Artificial Intelligence", and by Ariel Conn, "Artificial Intelligence: The Challenge to Keep It Safe", have a similar undertone.
It is also worth mentioning that Geoffrey Hinton, one of the pioneers of artificial intelligence and deep learning, has stated that "his breakthrough method should be dispensed with, and a new path to AI found". Quote from the article: "Speaking with Axios on the sidelines of an AI conference in Toronto on Wednesday, Hinton, a professor emeritus at the University of Toronto and a Google researcher, said he is now 'deeply suspicious' of back-propagation, the workhorse method that underlies most of the advances we are seeing in the AI field today, including the capacity to sort through photos and talk to Siri. 'My view is throw it all away and start again,' he said. [...] 'I don't think it's how the brain works,' he said. 'We clearly don't need all the labeled data.'" He also explained this view in a video talk. His own new approach to solving this problem is called "capsule network", and The New York Times featured an article about it: "A New Way for Machines to See, Taking Shape in Toronto". Medium also published a series about how Hinton's Capsule Networks work (part 1, part 2, part 3; a fourth part is still to come).
Links to Companies and Other Resources
An overview of companies investing in artificial intelligence is given by the articles "Artificial Intelligence Leaders You Must Know", "Where Major Corporations Like Salesforce And NVIDIA Are Investing In Artificial Intelligence" and "The Race For AI: Google, Baidu, Intel, Apple In A Rush To Grab Artificial Intelligence Startups". Regarding applications to neurosurgery, there is a systematic review that can be accessed for free. Concerning radiology, you might be interested in the article "The Future of Radiology and Artificial Intelligence".
Tons of links regarding artificial intelligence and machine learning are provided in the article "The Non-Technical Guide to Machine Learning & Artificial Intelligence". A list of courses for beginners can be found in the articles "How To Become A Machine Learning Engineer: Learning Path", "Over 150 of the Best Machine Learning, NLP, and Python Tutorials I've Found" and "My Curated List of AI and Machine Learning Resources from Around the Web".
It is also worth mentioning that the Nature Publishing Group is going to launch a journal on Machine Intelligence in January 2019, called "Nature Machine Intelligence".
Introductory Articles and Tutorials
While machine learning is already having an impact on the real world, a future technology that will have an even greater impact and "has the potential to change the world" if it works out is quantum computing, as it is "exponentially faster than digital computers". If you are new to this matter, you might also want to check out EdgyLabs' article "11 Facts That Help Explain the Quantum Computer" as well as the article "Quantum Computing explained" on Hackernoon and Gizmodo's "What the Hell Is a Quantum Computer and How Excited Should I Be".
One of the most exciting headlines from the past year regarding quantum computing has been: "An Entirely New Type of Quantum Computing Has Been Invented". Quote from the article: "Broadly speaking, there are currently a number of ways to make a quantum computer. Some take up less space, but tend to be incredibly complex. [...] Some tried and true ways to capture a qubit are to use standard atom-taming technology such as ion traps and optical tweezers that can hold onto particles long enough for their quantum states to be analysed. Others use circuits made of superconducting materials to detect quantum superpositions within the insanely slippery electrical currents. [...] Thanks to coding information in both the nucleus and electron of an atom, the new silicon qubit, which is being called a 'flip-flop qubit', can be controlled by electric signals, instead of magnetic ones. That means it can maintain quantum entanglement across a larger distance than ever before, making it cheaper and easier to build into a scalable computer."
Another article reported that "Scientists Achieve Direct Counterfactual Quantum Communication For The First Time". A further breakthrough was that scientists have used a newly found phase of matter (topological insulator) "in shrinking an electrical component called a circulator 1,000 times smaller", which is "super good news when it comes to squeezing more qubits into a small enough space" ("Physicists Just Invented an Essential Component Needed For Quantum Computers").
Programming and Working with Quantum Computers yourself
For those who would like to experiment with quantum computers, "D-Wave Systems and IBM are offering free access to their superconducting quantum computers in the cloud, along with access to their continually evolving programming tools". It is also interesting that Microsoft has developed a new programming language for quantum computers. As Singularity Hub writes, quantum computing might even "demand a whole new kind of programmer".
Haig Ferris, the co-founder of D-Wave (which, according to PC World, is the only company that is currently selling a quantum computer), suggests that in combination with machine learning, quantum computing will revolutionize cancer research. Quote: "'Whether it's brain research or cancer research, understanding and being able to model and learn from various ways you might design a drug to address a particular cancer this is going to be probably the most important application that you and I will benefit and notice,' said Farris. When queried about how soon we would begin to see the impact of the company's work, the D-Wave co-founder said we are already living in the quantum computing era. 'It's happening now,' he said. 'We've been working with some startup companies on some machine learning applications for drug design. There's probably things within the next three or four years that will have had an impact in drug design that nobody will be talking about but it will have, in fact, happened. Ten years from now you will probably be being treated with some quantum computing designed drug.'" The media also reported about a "record-breaking molecule simulation using a quantum processor, opening new promising paths to chemistry and materials science" conducted by IBM.
As some people are pretty excited about Bitcoin and the Blockchain technology, one should not forget to mention that quantum computing might be an "existential threat" to Blockchain technology. There is also an article on that issue at MIT Technology Review.
Links to Companies
To get an overview of companies that are into quantum computing, you might be interested in checking out EdgyLabs' article "11 Companies set for a Quantum Leap in Computing".
Other Non-Standard Approaches to Computing
In an article at phys.org, the authors write that "scientists have built tiny logic machines out of single atoms that operate completely differently than conventional logic devices do". Quote from the article: "Instead of relying on the binary switching paradigm like that used by transistors in today's computers, the new nanoscale logic machines physically simulate the problems and take advantage of the inherent randomness that governs the behavior of physical systems at the nanoscale - randomness that is usually considered a drawback."
According to another article, scientists can now store data in DNA with 100% accuracy. As a reporter writes on ZDNet, DNA can now store 215 petabytes of data per gram - Science writes in this context: "DNA could store all of the world's data in one room". Also, the media report that "Sudoku Hints at New Encoding Strategy for DNA Data Storage". It may also be possible to physically create a computer that "grows as it computes", a so-called nondeterministic universal Turing machine, using DNA molecules.
This is a huge field of modern medicine that attempts to regenerate organs either by injecting molecules into the body that make it regrow organs on its own or by creating new artificial organs and transplanting them. It also encompasses stem cells and tissue engineering.
According to the collected news items, progress in this field has been made in 2017 regarding heart, lung, kidney, bone, skin, female genitale tract and even the brain.
Molecular Biology, Genetics and Gene Therapy
This field, which some consider a bit controversial (see the articles "Eugenics 2.0: We're at the Dawn of Choosing Embryos by Health, Height, and More" and "Do We Have the Right to Edit the Genes of an Entire Species?"), is making tremendous progress thanks to the CRISPR gene editing technology. In 2017, the first human embryos were modified in the US using this technology. There are speculations that CRISPR might cure cancer, muscular dystrophy, diabetes, and even HIV.
Artificial Life and Synthetic Biology
This field, which is related to both biotechnology and computing, is still an emerging field. The subfield of Artificial Life has been existing for 30 years now and it has focused on simulating complete organisms or ecosystems in silico, i.e. on the computer. There is still a lot of space for development. Synthetic Biology was founded by J. Craig Venter and his team in 2010 when they synthesized the first man-made bacterium with modified DNA.
There have been a couple of highly interesting publications in this field, covered by news items such as "New organisms have been formed using the first ever 6-letter genetic code", "New Artificial Chromosomes Set Stage for First Complex Synthetic Genome" and "Quantum Artificial Life in an IBM Quantum Computer".
Neuroscience and Brain Research
This is a broad and massively diverse field, ranging from basic sciences to the treatment of neurological disorders such as Alzheimer's, Parkinson's, and ADHD. There have been many publications in the past year and it is difficult to determine which have been the most important ones; but probably the reports about technology to upload the contents of a brain to a computer or to download from the computer to the brain are among the most exciting ones (such as this one).
Psychiatry and Mental Health
This is a field in which I have worked myself, and I am especially delighted at the publication "The role of stress-regulation genes in moderating the association of stress and daily-life psychotic experiences" and the article "Depression May Be a Physical Illness Linked to Inflammation"; which confirm some of the hypotheses my own research group adhered to. In addition, there have been several other publications speculating on the cause of psychiatric disorders. One of them suggested that some psychiatric disorders might be a pH problem. Also, the debate whether schizophrenia is really a disease entity has continued.
Immunity and Infectious Diseases
There have been several interesting publications in this field, such as "Immune and Nerve Cells Work Together to Fight Gut Infections" and "Rejuvenating the Thymus to Prevent Age-related Diseases".
Endocrinology and Hormones
Some of the more interesting publications in this field are about gut bacteria, which are supposed to talk to the brain through cortisol and be able to trigger a gene that protects against type 1 diabetes. Regarding my own group's focus on the connection between steroidal hormones and the immune system, there has been a related publication about the possibility of estrogen stopping infection induced brain inflammation.
Oncology and Cancer
There have been numerous publications about cancer treatment, biomarkers and possible agents causing cancer, but nothing really revolutionary.
Longevity and Aging
Most of the articles that have been published in this field are more of speculative nature, but some have brought concrete results, such as "Fatty worms live longer, according to Stanford study" and "Eat fat, live longer? Mouse study shows a high fat diet increases longevity, strength".
Nutrition and Metabolism
2017 has brought us a couple of highly interesting inventions, such as edible water orbs that can help replace plastic bottles and a graphene-based sieve that turns seawater into drinking water.
This field has also brought some exciting news, such as "Nanochip could heal injuries or regrow organs with one touch" and "Nanoparticle-drug combo turns white fat to brown fat with potential to treat obesity, diabetes".
Physics and Chemistry
2017 has brought us a couple of exciting news in physics and chemistry, such as "Researchers create self-sustaining bacteria-fueled power cell" and "Gold chains give DNA semiconducting powers". There have also been inventions such as a spray that can turn any surface into a touch screen, a 2D magnet and a bionic lens that could push eyesight beyond 20/20 vision. Moreover, various websites reported that physicists confirmed evidence of a possible fifth fundamental force, that they discovered an "angel particle" that is its own antiparticle and that they demonstrated that 'impossible' tetraquarks can exist after all. Certainly some of these reports are to be taken with caution, but still, they are highly interesting and exciting.
In 2017, Artificial Intelligence has been the main buzzword. We have witnessed the success of AlphaGo Zero and can only wonder what is to come next year. Will Geoffrey Hinton's "Capsule Networks" cause the next paradigm shift? Is Ray Kurzweil right in that we will some day experience the "Singularity", when machine intelligence will surpass human intelligence?
Also, there has been considerable progress in Quantum Computing, threatening to render the Blockchain technology obsolete, which is currently quite popular considering the enormous increase of the value of a Bitcoin in the past few months.
In Biomedical Sciences, Gene Editing via CRISPR, Artificial Chromosomes and Brain-Computer Interfaces seem to be the hottest technologies where most progress is being made right now.
I am delighted that recent publications in Psychiatry, Endocrinology and Immunology have confirmed or at least supported some of the hypotheses that have been proposed by my own research group with Uwe Rohr.
Edible water orbs and graphene-based sieves have been some of the outstanding inventions of the past year, as well as bacteria-fueled power cells, a 2D magnet and other things I mentioned in the section about Physics and Chemistry.
All in all, it has been a decent year. See you again in 2018!