Here are the three winners of our The Time is Write 2.0 competition! You can read the three articles below, but first: congrats to Dipika Rajesh, Aditi Balaji and Pratyush Singh.
The Time is Write is an article writing competition that encourages all the aspiring writers to lay out their thoughts in writin and to share them on a global platform. This year, participants had to write a short article on the topic “Your Dream Software: Revolutionize the future”, about what their idea of a perfect software might be in order to revolutionize a particular field.
Asclepius Will See You Now by Dipika Rajesh (Sri Venkateswara College of Engineering)
In the 1960’s, Hanna-Barbera’s space-age show The Jetsons portrayed the future as a technological utopia—surrounded with quirky contraptions, robot companions, and, not to forget, the flying cars. Fast forward half a century, and we are in 2017, the age of cryptocurrency, social media, the Internet of Things, artificial intelligence, self-driving cars, and robotic voice assistants.
Technology has come far, seeing increasing developments in areas of intelligence, automation and data accuracy. However, some areas of recent advancements lack answers to critical questions. For example, self driving cars still have major security concerns about being hacked and medical AI lacks doctor-patient intimacy. With “The Future” seemingly within our reach, but still reasonably ahead of us, the revolutionary technologies in the making need to cater to social causes, which includes healthcare, in order to ensure a stable, sustainable future. To this end, many companies and research facilities, such as IBM and daVinci Surgery, have already begun, and in some cases finished, building artificially intelligent robotic technologies that assist in pertinent fields such as surgery and diagnostics. One possible reason behind the increased amount of attention towards the field of intelligent medicinal applications is the desire to eliminate miscalculations and misconceptions.
A study[1] shows that 12 million misdiagnoses occur annually in the US. Another such study[2] shows that more than 4,000 “wrong site errors” occurred yearly in the United States known as “never events.” A technology that aims to correct these errors should combine both the diagnostic aspects and the surgical precision that is required in healthcare.
The God of Medicine
Let’s imagine a cross between J.A.R.V.I.S from Marvel’s Iron Man and Baymax from Disney’s Big Hero 6—with more delicate and precise hands for surgical purposes, of course—, and let’s call it “Asclepius,” after the greek god of medicine.
All physicians are expected to be up to date with every diagnostic possibility to exist along with recent developments within the medicinal field they specialize in. This task of knowledge consumption can be very taxing for a human brain, but very simple for a state-of-the-art Asclepius with the required memory space. All the information required to run the diagnostics would be stored in Asclepius’s enhanced database, with the necessary faculties to make observations. In situations where doctors tend to use hackneyed methods to approach a prognosis, an intelligent machine would be able to provide insight with cold hearted objectivity. Therefore, Asclepius would be programmed to take up the responsibility of a diagnostician that is free of these biases, much like IBM’s cognitive question-answering system, “Watson”[3], which can be considered the best diagnostician currently. For example, in August 2016 it was able to give a proper and detailed diagnosis for a Japanese leukemia patient whose case had stumped doctors for months, thereby proving its efficiency.
Further, in the line of surgery, robot-aided minimally invasive surgeries have already been widely implemented across the globe. Currently, robots that assist the surgeons are either partially or completely controlled by trained officials with the help of a monitor to gain a magnified view. Total takeover by technology with Asclepius would result in enhanced vision, increased precision and more control over the entire operation, and with Asclepius, we could incorporate the knowledge of its diagnostics database directly into the surgical procedure. A complete state space search through the database should produce the most algorithmically optimal step forward. For example, in April 2016, the first Virtual Reality Surgery[4] was broadcasted by Shafi Ahmed, a cancer specialist and surgeon. This was achieved with the help of a 360°camera that live streamed every movement globally in VR. Asclepius could learn surgical procedures using the knowledge it gains from machine learning algorithms trained on videos and documentation of past procedures that together cover all recorded nooks and crannies of every possible surgery, mishap, and recovery.
Finally, to develop an interpersonal relationship with patients, apart from sentences that are directly processed by natural language processing, the patient could use a wearable to transmit bio-information to Asclepius wirelessly. Asclepius would then have the power to provide health checkups based on the patient’s past records and their current condition, upon which it would run proficient diagnostic searches. Then, to complete its role as a doctor, Asclepius would calculate dosages automatically based on aggregated data from the patient’s personal information. This medical prognosis, done by an intelligent machine, would be personally tailored, appear less formal, and still have accurate, precise results.
Conclusion
It has been said that there have been more monthly visits to WebMD’s website[5] and other health resource websites than doctors altogether. From this, we gather that people prefer information at the tip of their fingers with minimal effort and time taken without going through a formal conversation with a professional. A revolutionary software like Asclepius would provide real time diagnostics with minimal interaction, catering to this need of the public. And yet, with all that being said, currently it is only based on human intuition—which can only develop over time and with experience—that many lives are saved, every minute and against all odds. This level of importance of human intelligence is yet to be captured by machines.
At the end of they day patients look for a caring person to cure them of their troubles. Despite its effectiveness in delivering solutions, people prefer personal contact, something that Artificial Intelligence is still cracking—but not for too long by the looks of it.
If these obstacles are overcome, Asclepius will soon be less of a dream and more of a reality. With technology rapidly developing, every human function is becoming mimicked by intelligent, error-proof machines. For all the resources we take from society, we can give back only through technology that makes life easier. So, irrespective of how many apples you eat, this revolution is here to stay!
References
- “Outpatient Diagnostic Errors Affect 1 in 20 U.S Adults, AHRQ study finds”, published April 16, 2014, 2017,https://www.ahrq.gov/news/newsroom/press-releases/2014/diagnostic_errors.html
- “Never Events”, accessed January 14, 2017, https://psnet.ahrq.gov/primers/primer/3/never-events
- “IBM’s Watson gives proper diagnosis after doctors were stumped”, published August 07 2016,NY Daily News, http://www.nydailynews.com/news/world/ibm-watson-proper-diagnosis-doctors-stumped-article-1.2741857
- “What’s next for virtual reality surgery?”, published May 23, 2016,http://www.wired.co.uk/article/wired-health-virtual-reality-surgery-shafi-ahmed
- “Technology Will Replace Many Doctors, Lawyers, and Other Professionals”, published October 31, 2016, https://hbr.org/2016/10/robots-will-replace-doctors-lawyers-and-other-professionals?sf50989575=1
Dipika Rajesh is a third year Computer Science student at Sri Venkateswara College of Engineering, Chennai. Her love for animals is only superseded by her love for coding. She spends her free time learning about AI, reading books, playing the guitar and mastering a variety of board games.
Life is but a dream by Aditi Balaji (Sri Venkateswara College of Engineering)
As a child, I used to check under my bed every night before going to sleep. However, in a manner atypical of other six year old, I never checked for monsters – I checked for ninjas who might attack me during my sleep. At one point, my paranoia got so overwhelming that I had to check the entire house before I could let my parents turn me in for the night. Thankfully, over time, my paranoia faded away, letting me explore more appealing fantasies fueled by my rather fertile imagination.
That childhood phobia might explain why I have been fascinated by the subject of lucid dreaming since I found out about it, many years ago, from a YouTube video by ASAPScience [1]. Enthralled by the concept of the dreamer consciously recognizing that they’re in the throes of a dream and then exerting some measure of control over their actions in said dream, I spent the entire summer researching lucid dreaming. Of utter interest was the practical application of this controversial technique in the treatment of recurring nightmares.
One thing that both scholars and amateurs agree upon is that lucid dreaming isn’t as easy or quick as daydreaming. It involves great effort and harsh training of the mind. What I learned from Charlie Morley’s Dreams Of Awakening was that it is possible to create our own methods to actively be consciously sub-conscious during our dreams. Nevertheless, lucid dreaming is not an easy skill to learn: it takes time, patience and many more months, maybe even years before one can use it as a cure for haunting nightmares.
As a Computer Science student, even though my knowledge of psychology is basic, I couldn’t help but think that the strides we have made in the area of virtual reality in the recent past could bring the fruits of lucid dreaming to people who suffer daily – or rather, nightly. This was the motive behind my dream software, pun not intended. It would simulate the conditions of the dream that the victim often suffers through by using advanced graphics to re-construct the dream, and use virtual reality devices to administer it to patients when they are fully awake, in an attempt to help them conquer their fears when they are in full control of their faculties.
This software, the way I see it, has two crucial components: a highly immersive, interactive virtual reality system, and a fast learning AI, which takes in the patient’s initial description of the dream and generates a “scene”. The scene changes dynamically based on the patient’s reactions, and as the patient remembers more, the AI makes the relevant changes and introduces new stimuli into the dream. Thus, the patient truly interacts with a fantasy made tangible; an experience essentially imitating a lucid dream. And when the dream gets overwhelming, the patient can simply terminate the scene, just as an expert lucid dreamer can stop dreaming at the drop of a hat. The overall experience is entirely tailored for the user, and the software’s flexibility makes it suitable for therapy.
Extending the scope of this software, I believe it would also help in the treatment of many kinds of psychological diseases and phobias. Evolution, so far, has taught us to value fear, because it acts as a catalyst to instinct, which has ensured the survival of our species. But phobias are chronic, paralyzing, crippling fears which inhibit the victim’s ability to live life to the fullest. A few examples: agoraphobics cannot step out of their house, owing to their fear of open spaces. Most of them remain in the safety of the four walls of their houses, unable to seek the help they need, since that requires stepping out. People suffering from climacophobia, the fear of staircases, can get stuck on the top floor of a high rise building which is burning, unable to use the emergency staircases. Enochlophobia, the obsessive fear of crowds, doesn’t have a cure, but does get better with medical help; and while one might find silly the fear of pores and holes, trypophobia does actually touch some of us: a simple walk to the nearby park turns into a trip to the closest hospital because of a piece of sponge lying in a ditch.
While it is up to the brave individual who suffers from such conditions to decide how to deal with their phobia, this software could help victims of trauma heal their psychological wounds, and enable chronically paranoid people to come to terms with their paranoia. It could help them chart out a plan of action, a method to survive, and the virtual dreamscape could be manipulated to help the user create a “safe haven”. Additionally, the AI could help diagnose specific psychological problems and report them to a psychologist, and based on the treatments suggested, could integrate them into the therapy process.
Like any good computer scientist, I am the first one to admit that while my dream software might help revolutionize psychotherapy, feasibility is an issue, as of now. We do have a good idea about the mechanics of the eye and how images are re-constructed by the brain, but we don’t know enough about neural signals to correctly interpret and reproduce the complex patterns of phobias. However, regarding the great progress that has been made in the fields of brain analysis and neurosciences, I am confident that projects such as the organic artificial synapses developed at Stanford University [2], the AI virtual consultant used by UCLA’s radiology department to deliver better health care [3], and the new software to decode digital brain data created by researchers at Princeton University [4], will be the pebbles of impetus that set forth the avalanche of technological psychotherapy.
No matter how much we, as a species, claim to be of a higher thinking order compared to other living creatures, fear, and the brain’s response to it, is what ties us down to our primitive origins. Fear is the most basal human emotion, and has been instrumental to the survival of our species for so long. But in order to truly break barriers and reach new pinnacles in terms of evolution, in order to truly achieve greater sentience, we need to conquer our irrational fears and phobias, and I genuinely believe that my dream software has the potential to take evolutionary psychology and psychotherapy to new heights.
References:
- “The Science of Lucid Dreaming”: https://www.youtube.com/watch?v=lYSX51xBkos
- “Artificial synapse for neural networks”: https://www.sciencedaily.com/releases/2017/02/170221142046.htm
- “AI virtual consultant for better patient care”: https://www.sciencedaily.com/releases/2017/03/170308114842.htm
- “New software for decoding digital brain data”: https://www.sciencedaily.com/releases/2017/02/170224133923.htm
Aditi Balaji is a third-year Computer Science Engineering student at Sri Venkateswara College Of Engineering, Anna University, India. Besides data analytics and quantum computing, she is also interested in debating, running marathons and philosophy. She is a huge bibliophile and hopes to publish an original novel some day. She also plans to visit all the countries in the world at least once.
Your Dream Software: Revolutionize the Future by Pratyush Singh (Indian Institute of Information Technology, Allahabad)
Bioinformatics is an up and coming discipline with an insane amount of research potential. Ever since Information technology has helped develop methods and software for understanding biological data, the information we gather and analyse has exploded both in quantity and value. Bioinformatics applies computer science, statistics, mathematics, and engineering to examine and understand biological data.
Genomics and Proteomics are vital parts of modern bioinformatics. The ability to study genome and protein sequences has helped us unlock the answers to innumerable mysteries of nature[1]. Did you know that nearly 99.9% of you is genetically similar to any other person, and only about 0.1% of your genome sets you apart from about 7 billion other people?
After we learned how to sequence genomes, it begged the question – what would happen if we could somehow modify them, engineer them? This gave rise to a series of genetic engineering experiments where the goal was to make genetic modifications to pre-existing life forms through examination of genetic characteristics and the effect of their modification in the host organisms. Today, in thousands of laboratories across the world, there are researchers conducting experiments on lab rats and guinea pigs in order to determine the effects of gene splicing and insertion into these host organisms. One such example is the University of Tokyo, where researchers investigated a Damage suppressor (Dsup) protein in the Tardigrade genome to realize that it provided protection against DNA damage from X-Ray radiation[2].
But this fashion of research has led to several ethical questions that leave some within the scientific society and the larger community outside the scientific circles in a moral dilemma. Should we, at the cost of our fellow species, seek to advance unhindered exploration of the potential that genomics holds? The sense of gratitude that the researchers owe to the humble lab rat can be gauged at the [3] Institute of Cytology and Genetics in Novosibirsk, Russia where a six feet tall statue stands as a monument built to honour them.
So, say you were given a tool of imagination that could fashion any software you could think of. Wouldn’t it be amazing if it could fashion something to extinguish this ethical predicament? What if we could combine the research we have done so far and feed it into machine learning and then utilize probabilistic and statistical mathematics to create unbelievably accurate models of genomic combinations we have not done yet. What if we could visualize what effects would this eventually have on the host organism rather than having to use actual animals to study the same?
Background Information
Since we are talking about the future here let us assume that an unusually large amount of genetic research would have been done by the time someone got down to the task of creating such a software.
Machine learning and artificial intelligence have advanced to such an extent today where we are using them to create other intelligent systems. [4] In one experiment, researchers at the Google Brain artificial intelligence research group had software design a machine-learning system to take a test used to benchmark software that processes language. What it came up with surpassed previously published results from software designed by humans.
I can only imagine what these technologies will be capable of in the future. Now say we combine all the lab research we have done by that point in time and feed it into deep learning models which have learned how to learn, we could use these to accurately chart data that we need to model results of genetic experiments.
This background is rather important since even an artificial brain is a brain just like ours albeit a far more advanced one. It uses the available information to extrapolate and then predict outcomes with a greater accuracy than ours. I should go ahead and mention that I am assuming that these systems would have advanced to an extent where the results they produce should provide near 100% accuracy even though it is mathematically impossible.
The Software itself
This will be a technology that can be used to perform alteration to the genomic character of biological life forms, also comprising the relocation of the genetic factors within and across species to design fresh life forms.
It will give us the ability to interleave DNA sequences in the host genome by first segregating and replicating the genetic material we are interested in. Genes may be deleted or added. This will open up a host of possibilities apart from simple grafting.
Other major features would be visualization and automation. 3-dimensional projections will go a long way into helping picture the changes at various levels. Automation will help suggest genetic mixing which we haven’t even envisioned and probably lead to a plethora of other organisms.
The software should be a distributed system that has data gathered from systems all across the world. We would need middleware services combined with hybrid peer to peer models in order to facilitate a free and open exchange of data so that the software can function efficiently. Apart from this, we will also have better load balancing and fault tolerance which is always a big ask for highly complex systems such as these.
This technology can assist research in agriculture, medicine, defense, etc. Our world is changing at a far greater pace than natural selection can help most of us adapt to it. Going forward the debate concerning human intervention in natural selection will hold great importance as advancements in bioinformatics strive to keep up with the unprecedented changes in our environment. Such a technology will be instrumental in advancing the cause of the field without having to destroy a part of the environment in the process.
In the future statistics and probability can help us predict with near perfect accuracy what effects genetic modifications will manifest in organic life forms including humans. We can’t claim we are trying to preserve life and destroy it in the process of learning how to do so. Life, be it human or otherwise holds the same value to the natural balance. Hopefully, we will come to a point where we can extend that respect for life to all our fellow beings as well.
References
- http://www.nature.com/nature/journal/v491/n7422/full/nature11632.html
- http://www.nature.com/articles/ncomms12808
- http://boingboing.net/2014/01/17/a-monument-to-laboratory-rats.html
- https://www.technologyreview.com/s/603381/ai-software-learns-to-make-ai-software/
Pratyush Singh is a 3rd year Electronics and Communication (specialization: Biomedical Engineering) student at the Indian Institute of Information Technology, Allahabad, India. He is actively involved in research projects pertaining to VLSI devices (memristors, solar cells). He is also interested in writing, web development, distributed systems and graphic design.