Archive for the ‘Artificial Intelligence’ Category

Artificial Intelligence and Machine Learning for Insurance Technology from Johnson Controls Available on the Ocean Tomo Bid-Ask Market – Yahoo Finance

TipRanks

Today, were looking at two small-cap biotech firms whose stocks have struck a rut. Each company has hit a recent clinical setback that sent the share price falling, erasing previous gains and sending it back down to low levels. Setbacks of this sort are not uncommon in the biotech industry, and in fact highlight the risk and speculative nature of the industry. So what should investors do, when a stock collapses? Is this a matter of poor fundamentals? And has the stocks price found its low point yet? Thats where the Wall Street pros come in. Noting that each is set to take back off on an upward trajectory, some 5-star analysts see an attractive entry point for both. Using TipRanks database, we found out that these two tickers have earned Moderate or Strong Buy consensus ratings from the analyst community, and boast strong upside potential. Cortexyme, Inc. (CRTX) The first beaten-down name we're looking at is Cortexyme, a clinical-stage biopharma company focused on degenerative diseases, especially Alzheimers. The company's lead candidate is COR388, also called atuzaginstat. Atuzaginstat is currently under investigation in the GAIN trial, a study of its efficacy against Alzheimers disease. The trial is fully enrolled, with 643 patients, and the company was moving toward an open label enrollment (OLE) section of the Phase 2/3 study. During a routine regulatory update, Cortexyme announced that the OLE phase would be halted, although the primary GAIN study will continue, with results due to be released in Q4 2021. The announcement of the partial halt triggered a 35% drop in share price. The partial hold was prompted by adverse events on the liver during the atuzaginstat trial. The hepatic symptoms were reversible and showed no long-term lasting effects. The FDA reviewed these records, and in collaboration with Cortexyme the decision was made to hold the OLE while continuing with GAIN. This decision allows the main thrust of the program to continue, while working out a new protocol for the OLE. The purpose of the OLE is to test long-term efficacy and tolerability of the drug. In a review of Cortexyme after the announcement, HC Wainwrights 5-star analyst Andrew Fein noted, Cortexyme's announcement of a partial clinical hold on the OLE study of atuzaginstat is disappointing, but the reversible nature of the liver toxicity might provide some ray of hope for Cortexyme. We believe that the pivotal trial's continuation suggests that the drug-induced liver injury might not be severe enough to halt the program. Turning to the near-term, Fein adds, Continuation of the GAIN trial is encouraging despite the partial hold on OLE. It suggests that FDA plans to wait for the additional data from the pivotal trial before coming to any conclusion. Management shared that nearly one-third of the GAIN patients have completed the study and way past the 12-week time point, suggesting that they are out of risk. To this end, Fein rates CRTX a Buy, and his $76 price target indicates confidence in a 147% growth potential. (To watch Feins track record, click here) Overall, Cortexyme has a Moderate Buy rating from the analyst consensus, with 6 recent reviews breaking down 4 to 1 to 1, Buy-Hold-Sell. The stocks $83.60 average price target suggests that Wall Street sees a high potential here, on the order of ~170% upside from the trading price of $30.74. (See CRTX stock analysis on TipRanks) Immunovant (IMVT) Next up is Immunovant, a clinical stage biopharmaceutical research firm, focused on developing treatments for patients with autoimmune disorders, a class of diseases in which the immune system attacks the patients own body. The firms lead drug candidate, IMVT-1401, is undergoing trials as a treatment for thyroid eye disease, myasthenia gravis, and warm autoimmune hemolytic anemia. The drug described as a novel, fully human anti-FcRn monoclonal antibody, delivered by subcutaneous injection. On February 2, Immunovants stock plunged 42%, and it has been falling ever since. The precipitating factor was an announcement by the company that IMVT-1401 has had its Phase 2b clinical trial, for thyroid eye disease, halted temporarily, due to patients experiencing dangerous rises in their LDL levels. LDLs are the potentially harmful form of cholesterol, which have been connected to cardiovascular disease. Despite the clinical setback, Stiffels 5-star analyst Derek Archila reiterated a Buy rating on IMVT shares, along with a $28 price target. This figure suggests a 52% upside potential from current levels. (To watch Archilas track record, click here) Interestingly, increases have only been seen in TED patients, and our review of the literature suggests a few things: (1) it's likely this is TED specific given the biology- see below for details, but we don't think similar LDL increases will be seen in other indications outside TED; and (2) other anti-thyroid therapies used in Graves/TED also see similar increases in LDL, which end up being transient. We think IMVT-1401, in away, is replicating this mechanism," the analyst noted. Archila summed up, "While we will need to see additional data from the company to confirm... we don't think this program is dead. Overall, the Strong Buy analyst consensus view on IMVT would suggest that Wall Street generally agrees with Archilas assessment. This rating is derived from 8 recent reviews, which include 7 Buys and only a single Hold. The average price target here stands at $40.38, implying ~121% upside for the next 12 months. (See IMVT stock analysis on TipRanks) To find good ideas for stocks trading at attractive valuations, visit TipRanks Best Stocks to Buy, a newly launched tool that unites all of TipRanks equity insights. Disclaimer: The opinions expressed in this article are solely those of the featured analysts. The content is intended to be used for informational purposes only. It is very important to do your own analysis before making any investment.

See the article here:
Artificial Intelligence and Machine Learning for Insurance Technology from Johnson Controls Available on the Ocean Tomo Bid-Ask Market - Yahoo Finance

[Full text] Artificial Intelligence in Medical OSCEs: Reflections and Future Devel | AMEP – Dove Medical Press

Introduction

As the age of Artificial Intelligence (AI) begins, an increasing number of industries are exploring the applications and possible ramifications of this new era on current practice.1 In that sense, medicine and healthcare is the same; there have been numerous commentaries regarding the increasing role of AI in healthcare, with both proponents and opponents for the further application of AI and deep machine learning in clinical practice.2,3 Various arguments have been cited, with advocates citing the efficiency and accuracy of AI in easing the clinical workload, given the fact that network-based interfaces allow for knowledge and information to be shared across servers to allow for shared mastery of the fields programmed and learned by the machines.2,4,5 However, critics ruminate the possibility of a bleak future of employment in the healthcare industry and ethical dilemmas where responsibility for medical errors are sought.4,6 However, much of these remain conjecture, for there has been minimal research and study into the application of this rapidly progressing technology in medical practice.

While the future remains to be seen, ranging from a completely replaced healthcare workforce to one where AI plays but a supportive role of guiding clinical practice, or anywhere in between, AI would undeniably be playing a larger role in clinical practice in the future as compared to now.5 Therefore, the question ariseswhat skillsets and mindsets are required by clinicians in this new era, and are the current evaluation methods employed adequate in assessing the preparedness of our candidates?

Objective Structured Clinical Examinations (OSCEs) are a dominant assessment tool in healthcare education, allowing educators to assess practical performance reliably.7,8 The goal is to prepare medical students for practice-based learning and to train and test competence under standardized conditions.8 However, these standardized settings may artifically simplify the complexity of nonstandard, authentic patient encounters in real clinical environments.8 Clinical communication is reduced to tick-boxes on an examiners checklist, and learners simply strive to demonstrate behaviorsunder time pressure in pursuit of marks.8 OSCEs, in its current form, risk being just a barrier to gaining the title of doctor rather than of being able to truly assess their ability to practice as a doctor.8 Given the changing climate, the objective of education should and will need to refocus back to patient care with greater emphasis than before.8

AI applications in higher education can broadly be classified into four domains: profiling and prediction, assessment and evaluation, adaptive systems and personalization, and intelligent tutoring systems.9 Zawacki-Richer et al has shown in their systematic review that AI applications can perform assessment and evaluationwith great accuracy and efficiency.9 For example, Sanchez et al used an algorithm to match students to the professional competencies and capabilities required by companies, in order to ensure concordance between courses and industry needs.10 From perusal of the available literature, while many authors have discussed the diverse characteristics, skills, and knowledge with the advent of AI and the gaps in the current medical education framework,1115 few, if any, have discussed how exams will be influenced by developments in these technologies. Thus, we seek to consider the various possibilities of how AI could shape clinical examinations, specifically OSCEs, in terms of changes in design, curricular relevance, and methods of appraisal or assessment of these new skillsets, with or without the use of AI.

Demands on the modern clinician are ever-changing. While the clinicians of yesterday might have been expected to diagnose various complex conditions solely based off a simple consultation, examination, and rudimentary tests, clinicians of today have, at their disposal, advanced tools (including searches on a trusted medical information sites or phone applications) to aid their decision-making and diagnosis. Thus, the skillsets and problems faced by clinicians of different times vary greatly, and while some may be of the same flavor, these often take on very different contexts for the clinician to overcome. As the mercurial tides of medical reform are creeping up on us, the skillsets and demands on the clinician will continue to change.

In most places of the world, information flows more freely than ever before, with the increasing connectivity of the world and the decreased barriers to information. Gone were the days where the only source of information was from pouring over tomes of medical texts for a single piece of information; a simple search on a preferred search engine or literature site would likely yield answers to the question that was sought after. With the evolution of medical technology and advances in our understanding of medical conditions, the knowledge and information for all fields grow vastly, hopefully yielding more and more answers to the questions that were asked of yesterday and tomorrow. However, we live in paradoxical timesone where the ease of access to information does not necessitate greater knowledge, owing to the burgeoning knowledge base and limited human capacity. This limited human capacity is addressed by that of advancements in AI; an AI systems knowledge is limited not by capacity, for servers and storage space can be easily expanded, but by the progress of our own human knowledge, given that AI (in its current state for diagnostics) can only learn what is fed to it and draw logical deductions from the conclusions fed to it by human researchers and clinicians.2 This state of omniscience (of available knowledge) that AI possesses shows the advantage of AI over humans in terms of knowledge, and one that humans are unlikely to surpass.

This incongruity brings into focus one of the key changes in the skillset of the modern clinician: that of the increasing importance of correct knowledge capture over knowledge retention, as effectively discussed by Wartman and Combs.11 Given that a modern clinician is highly unlikely to surpass an AI system in terms of the amount of knowledge it drives, it is thus important that a clinician is broad-based, covering more breadth than depth. Of essence, however, is the ability to find reliable sources of information and knowing how to interpret and apply the information sought. This should thus prompt shifts in medical curricular planning to focus on the salient features of the above and to ensure that future clinicians have this skillset in their arsenal.

The implementation of AI into the medical field is an unstoppable force set to revolutionize the current landscape, for better or for worse. Thus, a new breed of clinician must be trained, one where doctors play a critical role in the design and planning of these systems and the direction that AI would evolve in, and in anticipating healthcare needs of the future. This necessitates knowledge in the domains of coding, big data, and user interface planning, which we believe would be core skills of the future. These clinical designers would combine clinical experience and knowledge and apply these principles in the design of AI systems which can then be extrapolated into more complex applications. The multidisciplinary team would also likely evolve to include computer or data scientists, to provide expertise on matters regarding these AI systems.

While design is one important aspect of these AI systems, another important facet would be the interpretation and application of what these AI systems provide; in other words, how can we make use of the predictions and recommendations of AI efficiently in clinical practice? This brings us to a core skill in the arsenal of the modern cliniciandata interpretation and translation into clinical practice. AI, in its current form, draws off past examples and the previously reached conclusions to structure and guide its future decisions. Decisions, diagnosis, and management are recommended based off the input of signs and symptoms of the informant, where that is the patient or the clinician. This, while infinitely helpful for common, recurring conditions, can prove to be our drawback in the event of new diseases and infections. While AI is able to determine new phenotypes and act as decision aids as to when to start resuscitation or certain supportive and life-saving managements, deep learning has a significant disadvantage in the way it functions; it requires data (large amounts of it due to clinical diversities, in fact) to draw conclusions before proceeding. Thus, we postulate that clinical acumen, contrary to what some might argue, is more important than ever, for while AI can generate more differentials (with confidence intervals) more accurately than a human, humans are still important to discern which patient has the common cold, and which needs to be isolated for a potentially species-threatening new virus. This is likely why AI is still unlikely to replace clinicians in the foreseeable future.

A modern clinician should, thus, have excellent clinical acumen, be able to discern and categorize the various complaints of a patient, and be able to utilize the aid of AI in making clinical decisions,16 without the algorithms replacing the clinician reasoning process.

To cure sometimes, to relieve often, and to comfort always. This central tenet (or dare I say, central dogma) of modern medicine is echoed by clinicians worldwide in medical education.17 This simple yet deeply profound saying epitomizes human touch; that of compassion and empathy and understanding of another human being in suffering. The human touch is unarguably essential in the field of medicine, which is often said to be both an art and a science. While efficiency and accuracy might arguably be improved by the implementation of AI systems, a dearth yet to be addressed has surfaced. AI systems have yet to, nor are expected to, fully replicate the human touch that clinicians can provide. While AI interfaces can offer simple lines expressing empathy or compassion, the absence of true emotion before these neural interfaces bequeath the main issuethat the human touch cannot be replaced by AI.18 This aspect of AI, considering its unstoppable nature of integration into medical care, has made communication skills and these soft skills of compassion and empathy ever more important, and is something that should be strongly honed by clinicians in this era of supposed replacement of roles by technologies and a shifting focus from the patient to the screen.19

The OSCE is a widely used clinical examination for the assessment of the clinical competency since its inception in 1975.7,20,21 Largely considered as the assessment of choice, the OSCE format has been modified in a multitude of ways to suit the syllabus and needs of each institution; for example, candidates at the National Taiwan University School of Medicine are assessed separately on each domain, whereas candidates at the National University of Singapore undergo what are known as Clinical Skills and Clinical Reasoning stations; in Clinical Skills stations, students are expected to perform a physical examination and generate possible differentials from physical signs, while in Clinical Reasoning stations, candidates are to take a history and perform a relevant examination before a discussion with the examiner about their differentials, most likely diagnosis, investigations, and management. Regardless of the format, the domains of history taking, physical examination, clinical discussion, and procedural skills often feature strongly and form the core of most, if not all, OSCEs. In keeping with the discussion above regarding the skillsets of the modern clinician, how then, will medical education, and specifically, OSCEs, evolve to bridge and meet the changing requirements of tomorrow?

An essential and indispensable skill of a clinician is that of history taking, as echoed in the aphorism by Hampton, A careful history will lead to the diagnosis 80% of the time.22,23 While the number presented might be arbitrary, there is certain truth in that sayingfor a detailed account of a patients presenting complaint and events leading to illness often gives critical clues and hints to the medical detective work that we are often tasked with doing.

In the age of AI, this skill is once again highlighted as something essential, for the use of AI interfaces to input presenting symptoms and complaints are restricted by several flaws, with 2 major ones that will be discussed here further. First, the AI system provides differentials and possibly a question list to ask from what is inputted into the system by the clinician. Systems like this would undoubtedly streamline processes and help doctors consider possibilities along that track, but this assumes that the original presenting complaint is interpreted correctly. The use of the algorithms of AI would potentially lock off the true diagnosis in the face of an error by a less astute clinician, which means that history taking is ever more important. Moreover, the algorithms of AI, while extremely powerful and robust, lose something in the process, that of the subtlety of patient complaints. While efforts can be made to attempt to differentiate the various symptoms from each other, a patient may not have thought something to be significant and reported to a front-facing interface of AI that it has a symptom. From the above, the need for good and effective history taking has been illustrated, and it is ever important to be able to pick up on the subtleties that patients may not have provided.

History taking assessment should, thus, be focused on evaluating these skills as set forth above. Perhaps, since AI systems (even at the lowest level) should be able to generate algorithms and lists of questions to lead questioning by the clinician, the focus might shift slightly from just clinching the various diagnosis and generation of differential diagnosis but place more emphasis on the differentiation of the various presenting symptoms and complaints, and to identify the subtleties that separate the entities from one another. These subtleties should also include when to decide when a patient might be malingering, or providing false symptoms, as well as when to consider a patients history as unclear testimony requiring further revision and clarification.

Moreover, assessment should also consider the situations and scenarios that future doctors might operate in, with AI systems on hand to assist in exams. For academic rigor, however, perhaps only the systems with the minimum capabilities will be provided, to ensure that enough competencies are achieved before allowing a student to progress on their journey into being a junior doctor. This also allows for assessors to gauge how well students know how to prioritize their lines of questioning in the limited time available; something that is already being tested now but will be even more important in the future given the immense amount of knowledge that AI is likely to have amassed. Thus, such a setup in assessment would allow for students to be proficient, independent of how AI might evolve where it could take on roles to complement clinicians as described in the minimum above, or in a larger role set to fully support clinicians with more robust systems to overcome the various operational bottlenecks we now face.

Another aspect worth considering would be the impact of AI on the conduct of such assessments. With advancement of technologies allowing for front-facing and interactive interfaces, these deep-learning systems present an opportunity for increased objectivity, cost efficiency, and standardization. Since these systems have a vast data network, clinicians and educators can set real-life cases to test students, while also ensuring standardization, for the AI system is one central network and would help reduce interpersonal variation. The use of such systems, coupled with other newer technologies such as Virtual Reality (VR) simulations,24 allow for many more students to take the exam at the same time, and would help save time and resources, as a one-time investment would save costs in the long run, compared to compensating patients for their time per exam.

In the domain of physical examination, while AI can aid in the integration of the history with the physical signs and the various possible differential diagnosis, we are of the view that a clinician is still required for a physical examination and assessment. While some might argue that a pan-scan of every patient would give all the information required, current capabilities are not able to provide an affordable, quick imaging method with acceptable levels of side effects, with the pricey MRI lacking the speed, and CT-scans exposing patients to large amounts of radiation. Thus, in this aspect, clinicians are still required to have good physical examination skills and to pick up the relevant signs before there can be much aid from AI systems.

OSCEs should, thus, still emphasize the need to pick up important and relevant physical signs; as with history taking, perhaps OSCEs can now be taken together with the aid of a minimal-assistance AI system, one which can suggest a physical examination that should be done in the presence of the previous history, and with some integration abilities after the reporting of the physical signs found. This would also provide an opportunity to test a candidates ability to interpret the integrated data and the suggested differentials before they can choose what they believe is the most likely diagnosis. In terms of OSCE conduct, front-facing interactive interfaces can be implemented into models with certain physical signs to aid testing and save resources.

Clinical discussions and viva questioning have been an essential part of most OSCEs, as they allow the examiner a glimpse of the train of thought of the candidates, providing an opportunity to gauge the abilities of each student.25 Traditionally, these questions have focused on the interpretation of various pieces of history and physical signs, as well as differential diagnosis and relevant investigations. Able students would also progress on to discussions about management of these patients.

However, in the age of AI, perhaps there would be a shift in emphasis in line with the shift of the knowledge-intensive era. This shift is likely to be that of a much stronger emphasis on approaches rather than conditions, where a candidates algorithm and ability to discern a condition from another might be more valuable than knowledge about a few conditions, for that is something that AI can provide with much less effort. While this is true, it does not give students an excuse to skive off without knowing about conditions, for basic information regarding each condition is still required to generate a sound clinical algorithm. This is especially crucial since one can walk down the wrong road with one wrong clinical judgment with AI, and it is of utmost importance that a clinician avoids this scenario, and where they already have made a mistake, be able to identify it early and quickly as well as set the clinical path on the correct track before any damage is done. This perhaps could be assessed as well; the signs, symptoms, and parameters that might suggest that something is suspected, as well as how to investigate in these scenarios.

In the realm of procedural skills, the use of AI could greatly enhance the efficiency and success rate of these repetitive, simple procedures.26 A medical student graduating to become a junior doctor is expected to be well versed in performing basic procedures, as they will be called to assist when there is difficulty in said procedures in patients in the ward. While there are convincing arguments that AI should be able to reliably complete these more menial tasks with higher success rates than humans, one must always consider the possibility that not all institutions and clinics might be equipped with such systems and capabilities. Thus, it is still an essential skill to know these basic procedures, such that in the event of AI interface failure or lack of such facilities at various institutions or clinics, the clinician is still able to reliably complete these procedures to continue the various steps in patient care.

When applying this concept to OSCEs then, perhaps one of the more drastic changes to this would be the marking scheme. In an exam, candidates are often graded on various administrative measures, such as certain steps for identification of patients, and other safeguards to prevent performing an incorrect procedure on the wrong patient, such that it is possible to pass the station even if they are unable to successfully complete the procedure. While essential to ensure patient safety, patient preparation and aftercare can easily be handled by AI and should not be the emphasis of such exams; perhaps, these procedural skills stations should place a heavier weightage on the success and completion of the procedure itself, such that our clinicians would be proficient at completing said procedures upon graduation. Perhaps, further assistance as with the minimal level required can be provided, such as image guidance for blood draws or the like. This would, however, be dependent on the resources available in each country and the minimum standard should be drawn from daily operational requirements of various hospitals.

OSCE conduct can also be greatly enhanced with AI, for it allows for a standardized patient to communicate with a student while completing said procedure required on a manikin or model. This might also help increase efficiency by reducing the number of examiners required for conducting exams in these stations. Moreover, AI can create simulated clinical environments for expansive learning, lessening environmental tension, and facilitate learners toward being fit to practice in the future. Pros and cons of AI applications in OSCE are summarized in Table 1. Musings from a future doctor (TKS) about the implementation of AI into healthcare proper are discussed in Supplement.

Table 1 Pros and Cons of AI Applications in OSCEs

Herein, we discussed the possibilities that AI might provide to OSCE examinations, which will reflect the changes in the skillsets that is required of the modern clinician. While AI still is in its infancy, we believe that it will play a pivotal role in the future, whether it be in healthcare or medical education and examinations, and while we do not claim to be prophets, we believe that this is the general direction toward which AI would lead medical education and hence its assessments. To stay relevant, we must continue to adapt and evolve as we navigate and overcome these uncharted territories that is our ever-evolving healthcare landscape; although we believe that there will always be a role for clinicians (albeit in different capacities), we must keep ourselves updated and be ready to accept and even influence change to stay as gatekeepers of these technologies. Given the theoretical nature of AI in OSCE (since the technology is still in its infancy), further study is required to further elucidate the role of AI in OSCE, and to the greater landscape of medical practice.

The authors thank Prof. Jann-Yuan Wang for providing critical comments.

The authors report no conflicts of interest in this work.

1. Alpaydin E. Introduction to Machine Learning. 4th ed. MIT press; 2020.

2. Rajkomar A, Dean J, Kohane I. Machine learning in medicine. N Engl J Med. 2019;380:13471358. doi:10.1056/NEJMra1814259

3. Sidey-Gibbons J, Sidey-Gibbons C. Machine learning in medicine: a practical introduction. BMC Med Res Methodol. 2019;19:64. doi:10.1186/s12874-019-0681-4

4. Shinners L, Aggar C, Grace S, et al. Exploring healthcare professionals understanding and experiences of artificial intelligence technology use in the delivery of healthcare: an integrative review. Health Informatics J. 2020;26:12251236. doi:10.1177/1460458219874641

5. Buch VH, Ahmed I, Maruthappu M. Artificial intelligence in medicine: current trends and future possibilities. Br J Gen Pract. 2018;68:143144. doi:10.3399/bjgp18X695213

6. Asan O, Bayrak AE, Choudhury A. Artificial intelligence and human trust in healthcare: focus on clinicians. J Med Internet Res. 2020;22:e15154. doi:10.2196/15154

7. Harden RM, Stevenson M, Downie WW, et al. Assessment of clinical competence using objective structured examination. Br Med J. 1975;1:447451. doi:10.1136/bmj.1.5955.447

8. Reid H, Gormley GJ, Dornan T, et al. Harnessing insights from an activity system - OSCEs past and present expanding future assessments. Med Teach. 2020:16. doi:10.1080/0142159X.2020.1795100.

9. Zawacki-Richter O, Marn VI, Bond M, et al. Systematic review of research on artificial intelligence applications in higher education where are the educators? Int J Educ Technol High Educ. 2019;16:39.

10. Snchez LE, Santos-Olmo A, lvarez E, et al. Development of an expert system for the evaluation of students curricula on the basis of competencies. Future Internet. 2016;8:22. doi:10.3390/fi8020022

11. Wartman S, Combs CD. Reimagining medical education in the age of AI. AMA J Ethics. 2019;21:E146152. doi:10.1001/amajethics.2019.146

12. Master K. Artificial intelligence in medical education. Med Teach. 2019;41:976980. doi:10.1080/0142159X.2019.1595557

13. Paranjape K, Schinkel M, Nannan Panday R, et al. Introducing artificial intelligence training in medical education. JMIR Med Educ. 2019;5:e16048. doi:10.2196/16048

14. Rampton V, Mittelman M, Goldhahn J. Implications of artificial intelligence for medical education. Lancet Digital Health. 2020;2:E111112. doi:10.1016/S2589-7500(20)30023-6

15. Wartman SA, Combs CD. Medical education must move from the information age to the age of artificial intelligence. Acad Med. 2018;93:11071109. doi:10.1097/ACM.0000000000002044

16. Loftus TJ, Upchurch GR Jr, Bihorac A. Use of artificial intelligence to represent emergent systems and augment surgical decision-making. JAMA Surg. 2019;154:791792. doi:10.1001/jamasurg.2019.1510

17. Taylor RB. Medical Wisdom and Doctoring: The Art of 21st Century Practice. Springer Science+Business Media, LLC; 2010.

18. Brown S. Preserving the human touch in medicine in a digital age. CMAJ. 2019;191:E622623. doi:10.1503/cmaj.109-5757

19. Alkureishi MA, Lee WW, Lyons M, et al. Impact of electronic medical record use on the patient-doctor relationship and communication: a systematic review. J Gen Intern Med. 2016;31:548560. doi:10.1007/s11606-015-3582-1

20. Harden RM, Gleeson FA. Assessment of clinical competence using an objective structured clinical examination (OSCE). Med Educ. 1979;13:3954. doi:10.1111/j.1365-2923.1979.tb00918.x

21. Harden RM. Revisiting Assessment of clinical competence using an objective structured clinical examination (OSCE). Med Educ. 2016;50:376379. doi:10.1111/medu.12801

22. Hampton JR, Harrison MJ, Mitchell JR, et al. Relative contributions of history-taking, physical examination, and laboratory investigation to diagnosis and management of medical outpatients. BMJ. 1975;2:486489. doi:10.1136/bmj.2.5969.486

23. Cooke G. A is for aphorism - is it true that a careful history will lead to the diagnosis 80% of the time? Aust Fam Physician. 2012;41:534.

24. Pottle J. Virtual reality and the transformation of medical education. Future Healthc J. 2019;6:181185. doi:10.7861/fhj.2019-0036

25. Hungerford C, Walter G, Cleary M. Clinical case reports and the viva voce: a valuable assessment tool, but not without anxiety. Clin Case Rep. 2015;3:12. doi:10.1002/ccr3.225

26. Hashimoto DA, Rosman G, Rus D, et al. Artificial intelligence in surgery: promises and perils. Ann Surg. 2018;268:7076. doi:10.1097/SLA.0000000000002693

See original here:
[Full text] Artificial Intelligence in Medical OSCEs: Reflections and Future Devel | AMEP - Dove Medical Press

The Most Powerful Artificial Intelligence Knows Nothing About Investing. That’s Perfectly Okay. – Institutional Investor

Theres no denying that 2020 was an exceptionally trying year. Few know this as well as active managers, who continued to struggle to provide the promised returns perhaps none more than celebrated quantitative investment managers like Renaissance Technologies, Two Sigma, and D.E. Shaw.

Clients have expressed growing disappointment with quants, as manifested in their steadily increasing redemptions.

Quant managers certainly recognize the peril they face. Some, like Ted Aronson of AJO Partners, have confessed that their investment models no longer work and have given up the ghost. In an interview with MarketWatch, Aronson pulled no punches for his decision to shutter his firm: Our return sucks over the past few years, he said. Our shit is so bad its unbelievable compared to our peers.

Others, perhaps lacking the stomach for such honest self-assessment, have chosen instead to selectively close only their woefully underperforming actively managed strategies (AQR Capital Management) or to remedy their ills by tweaking their current models (Bridgewater Associates). Both tactics offer the illusion of change but are more likely stalling tactics built on the hope that either client inertia or luck will allow them to extend their businesses.

However, a genuine remedy for their current and, most certainly, future ills exists: artificial intelligence.

Im not talking about machine learning (ML) techniques that quants and other managers have integrated into their investment processes over the past few years. Traditional ML techniques represent a significant expansion of the quantitative investors toolkit, but theyre not qualitatively distinct from traditional statistical methods, according to a white paper by Acadian Asset Management.

Used to augment traditional human quant models and methods, AI is seen as, at best, a handmaiden to human intelligence helpful, perhaps, but bound by the constraints of human intelligence.

However, the power of advanced AI such as deep learning (DL) and deep reinforcement learning (DRL) is rooted in its ability to find patterns in data directly and make predictions independent of human intelligence or expertise. While investment managers readily concede that these algorithms will solve incredibly complex problems in medicine, autonomous driving, engineering, robotics, and other verticals, they staunchly deny that DL and DRL will solve investment problems and build autonomous investment strategies.

That denial will be their downfall.

It is clear that their denial is based upon the single fundamental and universally held belief that investing is essentially and necessarily a human activity. Coldwater Economics Michael Taylor speaks for all of asset management when he writes:

This human-centric view of investing is such an integral part of the status quo that even those in the vanguard of ML quants like Jeff Shen, co-head of BlackRocks systematic active equity, are reluctant to envision a future in which humans are not central to the investment process. Fund management is an extraordinarily high-cognitive task, Shen told The New York Times last year. Were far away from turning on a computer and letting it run on its own.

Such a view of investing easily accommodates the use of traditional ML because, as Anne Tucker, faculty director of the Legal Analytics and Innovation Initiative at the Georgia State University College of Law, points out, this ML merely leverages components of human judgment at scale. Its not a replacement; its a tool for increasing the scale and the speed.

However, this view cannot accept AI that makes possible nonhuman investing, that autonomously learns and makes all the critical investment decisions and limits the role of humans to that of developers, not portfolio managers.

What is so challenging to incumbent investment managers is that this new wave of AI requires neither programming by humans to replicate the decision-making process of human experts nor deep domain knowledge of the disciplines in which it operates. Instead, through the use of deep neural networks, data, and computer power, this AI autonomously identifies in the data itself nonlinear statistical relationships undetectable to human-based and traditional ML methods.

For example, DL models used in cancer diagnosis and prognosis know nothing about medicine. Yet by focusing entirely on the data, they can achieve unprecedented accuracy, which is even higher than that of general statistical applications in oncology, according to a review published in the journal Cancer Letters.

It is the same with DL and DRL investing: These models know nothing about the investment canon, the CFA curriculum, value, or momentum. They are not programmed to mimic the decision-making of the greatest human investors. Instead, these algorithms hunt through the data, identifying patterns and similarities between the target and the data, and then use this knowledge to make investment predictions.

An instructive example of such powerful self-learning algorithms might be DeepMinds AlphaGo Zero, which was initially developed to play Go, the extremely complex Chinese board game that has more possible board positions than the number of atoms in the known, observable universe.

Unlike IBMs Deep Blue a human-designed, human-engineered, hard-coded computer program built in the 1990s to play the simpler game of chess AlphaGo Zero started tabula rasa, without human data or engineering and with no domain knowledge beyond the rules of the game. A DeepMind blog post from 2017 concluded that AlphaGo Zero used a novel form of reinforcement learning, in which AlphaGo Zero becomes its own teacher. The system starts off with a neural network that knows nothing about the game of Go. It then plays games against itself, by combining this neural network with a powerful search algorithm. As it plays, the neural network is tuned and updated to predict moves, as well as the eventual winner of the games.

Over the course of millions of games of self-play, the blog explained, the system progressively learned the game of Go from scratch, accumulating thousands of years of human knowledge during a period of just a few days. AlphaGo Zero also discovered new knowledge, developing unconventional strategies and creative new moves.

While AlphaGo Zero was specifically designed (but not programmed) to play Go, DeepMind reported that a later version of the program, AlphaZero, achieved similarly striking success with chess and shogi: AlphaZero quickly learns each game to become the strongest player in history for each, despite starting its training from random play, with no in-built domain knowledge but the basic rules of the game.

The unprecedented success of these and other experiments led the DeepMind team to draw the general conclusion that reinforcement learning can be used to achieve superhuman results in other domains:

In the face of this peer-reviewed and highly cited research (and other similarly robust research), investment managers continue to staunchly claim that humans will always . . . be a necessary component in the investment process.

This is the exact point in the argument where good quants should provide an abundance of empirical evidence in support of their claim of the impossibility of nonhuman investing.

Yet none is offered. Instead, they rest their case on suppositions, appeals to tradition, and strawman arguments.

Machine learning wont crack the stock market.

Critics using this meme generally imply that DL and DRL will not achieve the same 99 percent predictive accuracy these systems achieve in computer vision, speech recognition, and even cancer detection.

Such a level of prediction accuracy in investing would immediately conjure up memories of Bernie Madoff. What these critics fail to point out is that in investing, we are not trying to crack the code. Rather, as William Cohan writes in Fast Company, An exceptional trader would be thrilled with a 51 percent success rate similar to the house edge at a Las Vegas blackjack table.

Financial markets are not stationary. They change all the time, driven by political, social, economic, or natural events.

Denialists claim that, unlike MRIs and board games, markets are simply too complex and random and this complexity and randomness overwhelm the ability of DL and DRL to consistently find actionable insights in the data.

Putting aside the fact that there is no reason to believe human intelligence-based models could account for this complexity and randomness any better than any AI, it is important to acknowledge that complexity and randomness are human concepts derived from observations made by human intelligence of price activity, and that phenomena humans perceive as complex and random might not be perceived as such by DL and DRL systems. As Tom Simonite writes in Wired, Artificial intelligence is alien intelligence, perceiving and processing the world in ways fundamentally different from the way we do.

Jeff Glickman, co-founder of AI-based investment manager J4 Capital, makes this argument in Cohans Fast Company article:

But that doesnt mean that some other intelligence wont. Despite the fact that you or I might perceive it as being random, theres nothing random about it, he continues. Theres just an overwhelming amount of complexity thats beyond comprehension for humans, but within the ability of a massive supercomputer to comprehend.

And while AI is great at recognizing patterns in data and identifying similar situations observed in the past, it is usually at a loss on how best to act in new and previously unseen situations, such as the Covid-19 outbreak.

It is true that DL and DRL are trained on historical data and that circumstances not seen in the training data could prove troublesome.

Even a technology-heavy investment manager like Renaissance Technologies admits this. Bloomberg reports that Renaissance told clients in a September letter, It is not surprising that our funds, which depend on models that are trained on historical data, should perform abnormally (either for the better or for the worse) in a year that is anything but normal by historical standards.

But such an admission does not disqualify DL and DRL as possible investment systems.

If it did, it would also disqualify all human intelligence-based investment methods because they too use historical inputs (e.g., data, human experience) as model inputs and are at a loss on how best to act. (Quants rely on data from time periods that have no reflection of todays environment. When you have volatility in markets, it makes it extremely difficult for them to catch anything because they get whipsawed back and forth, Adam Taback, chief investment officer of Wells Fargo Private Wealth Management, told Bloomberg in November.)

[W]hile the breadth of data that can be used in finance is quite large, the time series are often very short and usually limited to a few decades. A limited number of time series observations means that any model using the data is also constrained to be proportionally small.

There is no question that DL and DRL are capable of ingesting massive amounts of data, and because of their large capacity, more data generally results in higher prediction accuracy.

Defenders of the status quo often point to the large data sets required to train computer vision models or autonomous vehicles and to the fact that financial market data is smaller. This limitation, therefore, disqualifies advanced AI from use.

Financial data might be limited, but the criticism fails to consider that, unlike traditional quant models, DL and DRL models are capable of ingesting nonfinancial data, including nontraditional data (e.g., geospatial data). And the type, kind, and volume of this data is growing daily, giving engineers a broader palette from which to paint. As my friend Chris Schelling points out:

Perhaps more importantly, those who disqualify DL and DRL from investing because of the perceived limitations of financial data offer no empirical evidence to support their claim. Because of the sparseness of input data (and not other factors), they cannot point to the live track record of a DL- or DRL-based investment strategy that has failed.

When it comes to applying AI [to investing], it is compulsory for us to understand exactly how each algorithm works.

Finally, we reach the black box objection, the trump card that deniers play if all other objections fail; the last stand of the status quo.

The thing about this objection is that, unlike the others, it is true. By its very nature, advanced AI is a black box. While we can observe how AlphaGo Zero plays, even its designers cannot explain why it makes a specific move at a specific time. Similarly, a manager using DL or DRL is able to provide a general overview of its approach (e.g., we use a recursive neural network) and can provide the models inputs and outputs, but it cannot explain why it makes a specific investment decision.

There are obvious counterarguments to this requirement of explainability. For example, the blackness of an investment model is the product of a specific historical epoch; option pricing models, technical analysis, program trading, optimization programs, and statistical arbitrage programs were the black boxes of their day. Others point out that we hold AI to a higher standard of interpretability than we do human intelligence because it is not possible to explain the why of human decision-making. As Vijay Pande, general partner at Andreessen Horowitz and former director of the biophysics program at Stanford University, wrote in The New York Times, Human intelligence itself is and always has been a black box.

The investment industry has made a clear choice.

It has chosen why over what, explainability over accuracy.

But this choice was inescapable because it preserves investing as a human activity.

This choice both accommodates the use of AI techniques that are not qualitatively distinct from traditional statistical methods and disqualifies the use of AI systems like DL and DRL that learn, make decisions, and take actions autonomously.

Above all, it ensures the quantitative investment process will remain recognizable in the foreseeable future, in Acadians words, thereby preserving the status quo. And the status quo is not working for many active managers, especially quants.

This choice should alarm asset allocators for the simple fact that it all but guarantees the perpetuation of the cycle of manager underperformance.

It is critical that allocators realize this future is the result of managers explicit decision to subjugate artificial intelligence to human intelligence and abrogate advanced AI, a choice grounded in hubris and self-interest and supported by strawman arguments, misunderstandings of advanced AI, and scant empirical evidence.

AI investing can help break this cycle of poor performance, but it requires that allocators choose what over why and accept, as Selmer Bringsjord toldVice, that we are heading into a black future, full of black boxes.

Follow this link:
The Most Powerful Artificial Intelligence Knows Nothing About Investing. That's Perfectly Okay. - Institutional Investor

Global Automotive Artificial Intelligence Market (2020 to 2026) – by Component, Technology, Process, Application, Vehicle Type, Demand Category,…

DUBLIN, Feb. 15, 2021 /PRNewswire/ -- The "Global Automotive Artificial Intelligence Market By Component (Hardware, Software, Service), By Technology, By Process, By Application, By Vehicle Type, By Demand Category, By Company, By Region, Forecast & Opportunities, 2026" report has been added to ResearchAndMarkets.com's offering.

The Global Automotive Artificial Intelligence Market is expected to grow at a steady rate during the forecast period. The Global Automotive Artificial Intelligence Market is driven by the growing adoption of advanced automotive solutions such as advanced driver assistance system (ADAS), adaptive cruise control (ACC), blind sport alert, among others by different OEMs. Additionally, government regulations to improve the safety in vehicles while assuring environmental sustainability is further expected to propel the market. Furthermore, ongoing technological advancements and new product launches by the major OEMs operating in the market is expected to create lucrative opportunities for the market growth through 2026. However, lack of proper automotive IT infrastructure especially in the emerging countries can hamper the market growth during the forecast period. Besides, high procurement and operational costs further restricts the market growth over the next few years.

The Global Automotive Artificial Intelligence Market is segmented based on component, technology, process, application, vehicle type, demand category, company and region. Based on technology, the market can be categorized into deep learning, machine learning, context awareness, computer vision, natural language processing and others. The deep learning segment is expected to register the highest CAGR in the market during the forecast period on account of the increasing popularity and adoption of self-driving cars.

Additionally, the major players operating in the market are also investing a lot in the development of self-driving cars. While the computer vision segment is also expected to grow significantly on account of its use in autonomous vehicles for signal recognition, image recognition, driver monitoring, among others. Based on process, the market can be fragmented into signal recognition, image recognition and data mining. The data mining segment is expected to dominate the market owing to the large volumes of data being generated and processed in autonomous and semi-autonomous vehicles. Based on application, the market can be grouped into human-machine interface, semi-autonomous driving and autonomous driving. The human-machine interface segment is expected to dominate the market owing to the growing need for providing enhanced customer experience.

Regionally, the automotive artificial intelligence market has been segmented into various regions including Asia-Pacific, North America, South America, Europe, and Middle East & Africa. Among these regions, Asia Pacific is expected to register the highest growth in the overall automotive artificial intelligence market owing to the growing demand for premium vehicle and increased adoption of AI and AI based services and solutions especially among the autonomous and semi-autonomous vehicles in the region.

The major players operating in the automotive artificial intelligence market are NVIDIA Corporation, Alphabet Inc., Intel Corporation, IBM Corporation, Microsoft Corporation, Harman International Industries Inc., Xilinx Inc., Qualcomm Inc., Tesla Inc., Volvo Car Corporation and others. Major companies are developing advanced technologies and launching new services in order to stay competitive in the market. Other competitive strategies include mergers & acquisitions and new service developments.

Objective of the Study:

The publisher performed both primary as well as exhaustive secondary research for this study. Initially, the publisher sourced a list of service providers across the globe. Subsequently, the publisher conducted primary research surveys with the identified companies. While interviewing, the respondents were also enquired about their competitors. Through this technique, the publisher could include the service providers which could not be identified due to the limitations of secondary research. The publisher analyzed the service providers, distribution channels and presence of all major players across the globe.

The publisher calculated the market size of the Global Automotive Artificial Intelligence Market by using a bottom-up approach, wherein data for various end-user segments was recorded and forecast for the future years. The publisher sourced these values from the industry experts and company representatives and externally validated through analyzing historical data of these product types and applications for getting an appropriate, overall market size. Various secondary sources such as company websites, news articles, press releases, company annual reports, investor presentations and financial reports were also studied.

Key Target Audience:

The study is useful in providing answers to several critical questions that are important for the industry stakeholders such as service providers, suppliers and partners, end-users, etc., besides allowing them in strategizing investments and capitalizing on the market opportunities.

Key Topics Covered:

1. Product Overview

2. Research Methodology

3. Impact of COVID-19 on Global Automotive Artificial Intelligence

4. Executive Summary

5. Voice of Customer

6. Global Automotive Artificial Intelligence Market Outlook6.1. Market Size & Forecast6.1.1. By Value6.2. Market Share & Forecast6.2.1. By Component (Hardware, Software, Service)6.2.2. By Technology (Deep Learning, Machine Learning, Context Awareness, Computer Vision, Natural Language Processing, Others)6.2.3. By Process (Signal Recognition, Image Recognition, Data Mining)6.2.4. By Application (Human-Machine Interface, Semi-autonomous Driving, Autonomous Driving)6.2.5. By Vehicle Type (Passenger Cars v/s Commercial Vehicles)6.2.6. By Demand Category (OEM v/s Aftermarket)6.2.7. By Company (2020)6.2.8. By Region6.3. Product Market Map

7. Asia-Pacific Automotive Artificial Intelligence Market Outlook7.1. Market Size & Forecast7.2. Market Share & Forecast7.3. Asia-Pacific: Country Analysis

8. Europe Automotive Artificial Intelligence Market Outlook8.1. Market Size & Forecast8.2. Market Share & Forecast8.3. Europe: Country Analysis

9. North America Automotive Artificial Intelligence Market Outlook9.1. Market Size & Forecast9.2. Market Share & Forecast9.3. North America: Country Analysis

10. South America Automotive Artificial Intelligence Market Outlook10.1. Market Size & Forecast10.2. Market Share & Forecast10.3. South America: Country Analysis

11. Middle East and Africa Automotive Artificial Intelligence Market Outlook11.1. Market Size & Forecast11.2. Market Share & Forecast11.3. MEA: Country Analysis

12. Market Dynamics12.1. Drivers12.2. Challenges

13. Market Trends & Developments

14. Competitive Landscape14.1. NVIDIA Corporation14.2. Alphabet Inc.14.3. Intel Corporation14.4. IBM Corporation14.5. Microsoft Corporation14.6. Harman International Industries Inc.14.7. Xilinx Inc.14.8. Qualcomm Inc.14.9. Tesla Inc.14.10. Volvo Car Corporation

15. Strategic Recommendations

16. About the Publisher & Disclaimer

For more information about this report visit https://www.researchandmarkets.com/r/2d457n

Media Contact:

Research and Markets Laura Wood, Senior Manager [emailprotected]

For E.S.T Office Hours Call +1-917-300-0470 For U.S./CAN Toll Free Call +1-800-526-8630 For GMT Office Hours Call +353-1-416-8900

U.S. Fax: 646-607-1904 Fax (outside U.S.): +353-1-481-1716

SOURCE Research and Markets

http://www.researchandmarkets.com

Read this article:
Global Automotive Artificial Intelligence Market (2020 to 2026) - by Component, Technology, Process, Application, Vehicle Type, Demand Category,...

Calastone implements Opsmatix Artificial Intelligence solution Improving client handling and efficiency within client operations – PRNewswire

LONDON, Feb. 16, 2021 /PRNewswire/ -- Opsmatix, an innovative provider of AI-powered omnichannel operations automationsolutions,today announcesthat Calastone, the world'slargest global funds network, has implemented the Opsmatix SaaS platformto process increasing business and email volumes into Calastone'sOperations Team.

Currently, Calastone supports some 2,500 clients in 44 countries and territories and processes over 200 billion of investment value every month. Opsmatix was selected following a rigorous proof of concept that demonstrated the system'sunrivalled automation capabilities in terms of categorising and understanding the intent of incoming client queries. The new system will enable the firm to scale itsclient handling capability as the firm growswhilst continuing to improve the clientexperience.This new approach reduces manual interaction on time-consuming tasks allowingthem to focus on more productive activities.

"We pride ourselves on providing aworld-class support serviceto our clients and look to how we can leverage the best technologies to drive continuous improvement,"says Mike Davies, Calastone'sGlobal Head of Operations.Opsmatix allows us to streamline the workflow management within the team enabling greater operational leverage and ultimately enhancing the overall client experience.Crucially we gain a much-improved system to manage workflow, together with an elegant case management user interface which enables us to categorise, escalate and manage any production issues in a more rigorous manner."

Justin Forrest, CEO at Opsmatix concluded. "We are delighted to be working and partnering with a customer of the calibre of Calastone. This relationship demonstrates Opsmatix'scapabilities and validates the many benefits the solution will deliver to the financial services sector and cross-industry. AI has come of age and is now a business imperative for all corporate operational functions using omnichannel communications involving unstructured data.Our goal is to be at the forefront of technology innovation and corporate advancement, and we are confident that Opsmatix has a pivotal partto play."

About Calastone

Calastone is the largest global funds network, connecting the world's leading financial organisations.

Our mission is to help the funds industry transform by creating innovative new ways to automate and digitalise the global investment funds marketplace, reducing frictional costs and lowering operational risk to the benefit of all. Through this, we generate the opportunity for the industry to deliver greater value back to the end investor.

Over 2,500 clients in 44 countries and territories benefit from Calastone's services, processing 200 billion of investment value each month.

Calastone is headquartered in London and has offices in Luxembourg, Hong Kong, Taipei, Singapore, New York, Milan and Sydney.

About Opsmatix

Opsmatix applies Artificial Intelligence (AI) to automate business communications and processes. It improves efficiency & quality, reduces repetitive tasks and accelerates operations based on multi-lingual long-chain omnichannel communications involving unstructured data and processes which require significant human intervention. Applications range from front-line customer service staff, contact centres and customer onboarding to manually intensive communications in the back office, including logistics and fulfilment. The OpsmatixSaaS platformsignificantly reduces the requirement for the wholesale offshoring of operational processing and call centres. The company was founded in 2018 and is basedin London.

Contact us via our website athttps://www.opsmatix.com/

SOURCE Opsmatix

Go here to read the rest:
Calastone implements Opsmatix Artificial Intelligence solution Improving client handling and efficiency within client operations - PRNewswire