Transductive conformal prediction addresses the simultaneous prediction for multiple data points. Given a desired confidence level, the objective is to construct a prediction set that includes the true outcomes with the prescribed confidence. We demonstrate a fundamental trade-off between confidence and efficiency in transductive methods, where efficiency is measured by the size of the prediction sets. Specifically, we derive a strict finite-sample bound showing that any non-trivial confidence level leads to exponential growth in prediction set size for data with inherent uncertainty. The exponent scales linearly with the number of samples and is proportional to the conditional entropy of the data. Additionally, the bound includes a second-order term, dispersion, defined as the variance of the log conditional probability distribution. We show that this bound is achievable in an idealized setting. Finally, we examine a special case of transductive prediction where all test data points share the same label. We show that this scenario reduces to the hypothesis testing problem with empirically observed statistics and provide an asymptotically optimal confidence predictor, along with an analysis of the error exponent.

One of the most frequently-used phrases at business events these days is "the future of work." It's increasingly clear that artificial intelligence and other new technologies will bring substantial changes in work tasks and business processes. But while these changes are predicted for the future, they're already present in many organizations for many different jobs. The job and incumbents described below are an example of this phenomenon. Steve Miller of Singapore Management University and I are collaborating on these stories.

F. Tan, and Ashish Khisti Abstract Motivated by real-world machine learning applications, we consider a statistical classification task in a sequential setting where test samples arrive sequentially. In addition, the generating distributions are unknown and only a set of empirically sampled sequences are available to a decision maker. The decision maker is tasked to classify a test sequence which is known to be generated according to either one of the distributions. In particular, for the binary case, the decision maker wishes to perform the classification task with minimum number of the test samples, so, at each step, she declares that either hypothesis 1 is true, hypothesis 2 is true, or she requests for an additional test sample. We propose a classifier and analyze the type-I and type-II error probabilities. We demonstrate the significant advantage of our sequential scheme compared to an existing non-sequential classifier proposed by Gutman. Finally, we extend our setup and results to the multi-class classification scenario and again demonstrate that the variable-length nature of the problem affords significant advantages as one can achieve the same set of exponents as Gutman's fixed-length setting but without having the rejection option. Index T erms Sequential classification, Empirically sampled sequences, Error exponents, V ariable-length I. I NTRODUCTION Quick and accurate classification is crucial in many real-life applications. For instance, to diagnose haematologic diseases based on blood test results, a physician wishes to detect the pattern, deviations, and relations in the blood samples of a patient as quickly as possible to make treatment plans. Similar challenges can be found in a broad range of applications such as genomics analysis, finance, and abnormal detection where there is an inherent tradeoff between speed and accuracy. In many real-world applications, classical hypothesis testing is infeasible due to the fact that the probability distributions of the sources are unknown. In practice, one often encounters classification problems in which one has access to training samples and is required to classify a set of test samples according to which distribution this set is generated from. To incorporate the real-life requirement of classifying the test samples as quickly as possible, one can consider the sequential statistical classification setup. This setup addresses the problem of classifying test samples given training samples with the additional requirement that the decision maker is required to make his/her decision based on as few tests samples as possible; it is however, known that all the test samples originate from the same distribution. The problem of classification using empirically observed statistics has been studied in many prior works.

Amazon Alexa users with health issues can now have Alexa call Healthtap's Doctor AI to help figure out what's wrong and direct them to act accordingly. Doctor AI initially supported Apple iOS and Android devices and its first voice application, Talk to Docs launched for those same devices in 2013. But support for Alexa, the smart voice-activated software that debuted with the Amazon Echo connected speaker, brings Dr. AI to home users who might not be adept at using screens. "We'd been doing text and video before, then expanded into voice and that's exciting in healthcare because we serve many populations that are older, disabled, or frail," said Ron Gutman, founder and chief executive of Palo Alto, California-based digital health company which claims 107,000 doctors in its network. Records and data from those doctors make up Dr. AI's health care data trove.

Now, if you need to talk to a doctor, you don't need to leave the comfort of your home. Amazon's (AMZN) Alexa is no longer just for checking the weather, playing music or shopping online. With a new app, you can now talk to Alexa about any health issues you may have. The Dr. A.I. app accesses data the company has collected over the years to help provide users with health information. Any time that you actually provide a set of symptoms in the context it will take the relevant data point, in this case it can be thousands, it can be tens of thousands of doctors' opinions that came over the years," Gutman told the FOX Business Network's Maria Bartiromo. According to Gutman, the system is set up to continually review and improve the information provided to users. "Not only answering questions, but also peer reviewing each other's answers for quality, right.

Digital health platform provider HealthTap is betting its new Dr. A.I. mobile app will eliminate the risk of a patient incorrectly self-diagnosing their condition through online searches by providing accurate, online triage that directs patients to the right level of care. The app, which uses artificial intelligence to perform online triage based on a patient's symptoms, can help reduce the risk of a patient incorrectly self-diagnosing her symptoms, which happens frequently with Internet searches, HealthTap says. There are 10 billion symptom-related health searches per year on Google, says HealthTap CEO Ron Gutman. "Search engines can't consult a patient's health records or ask follow-up questions to put a person's symptoms into proper context," Gutman says. "Effective triage requires detailed knowledge of a patient's personal health situation, making context critical to providing optimal care."

"My vision is to give people immediate gratification in healthcare," says Ron Gutman the high-energy and affable founder and CEO of HealthTap, a mobile health platform that connects customers with trusted health information and doctors in near real-time at any given time of day or day of the week from a network of over 50,000 U.S.-licensed doctors. "We call people'patients' but they are anything but patient when they are in pain or desperate to talk to a doctor during the day or in the middle of the night," continues Gutman. In creating a customer–focused health information service, Gutman sees the opportunity to build the world's first global mobile health brand. There is nothing small about Ron's ambitions or business opportunity. HealthTap looks to disrupt the model and bring high-quality information directly to the customer, instantly through their mobile devices for free or very modest costs.

If you've ever traversed the morasses of WebMD to try and figure out if that pain in your arm might signify a more serious medical condition, then you know just how hard it can be to pinpoint the most likely diagnosis (and its level of urgency) amid a sea of information. Digital health firm HealthTap wants to overcome that dilemma with its newest mobile app, Dr. A.I., which launched on Tuesday. The platform digitizes health care triage, HealthTap founder and CEO Ron Gutman tells Fortune, which is the process of assessing the level of medical risk facing a patient and the first step in the treatment pathway. Click here to subscribe to Brainstorm Health Daily, our brand new newsletter about health innovations. Gutman says that the problem with using conventional search engines to look up medical symptoms is that it's ineffective since they produce purely semantic results.

Every year billions of people visit sites like WebMD or NIH to find health information. However, a company called HealthTap wants to help you take your at home health care to the next level. HealthTap uses Artificial Intelligence and computer technology to give personalized recommendations on how to treat your symptoms, illnesses, and even tells you when it's time to see a doctor in person. HealthTap CEO Ron Gutman discussed how his company incorporated Artificial Intelligence into its mission to improve the standard of healthcare. "We created in the background this very valuable knowledge base and training set to help us now understand how to answer peoples' questions and direct them to the right level of care. So basically we're using Artificial Intelligence and machine learning to take this data and direct people to the right care, at the right time, at the right cost," he said.