When Doron Hazan '21 was drafted into the Israeli Defense Forces (IDF) after high school, he had the opportunity to join the army's intelligence unit. It was the obvious choice for the self-described "math and physics nerd" from Kiryat Shmona, a small town in Israel's Hula Valley just south of the Lebanese border. But Hazan was not one to make obvious choices. "All of my life I've been interested in human behavior," says Hazan, a junior who is enrolled in one of MIT's newest majors: computation and cognition, or Course 6-9. Launched in the fall of 2019, Course 6-9 is a joint curriculum offered by the Department of Electrical Engineering and Computer Science (EECS) and the Department of Brain and Cognitive Sciences (BCS).

Speech analytics technologies are used to extract information at customer contact points across various channels such as voice, chat, email, social channels, and surveys. Across the world, voice and phone interaction is the most common mode of communication used by consumers. Therefore, speech analytics is used in Voice User Interface (VUI) to derive insights at different contact points. In current times, organizations across various industry sectors are undertaking programs for transcripting and analyzing customer and organizational media. This is mainly to take logical decisions for customer and business management with the help of speech and text intelligence.

In the summer of 2009, the Israeli neuroscientist Henry Markram strode onto the TED stage in Oxford, England, and made an immodest proposal: Within a decade, he said, he and his colleagues would build a complete simulation of the human brain inside a supercomputer. They'd already spent years mapping the cells in the neocortex, the supposed seat of thought and perception. "It's a bit like going and cataloging a piece of the rain forest," Markram explained. "How many trees does it have? What shapes are the trees?"

Deep reinforcement learning techniques have shown to be a promising path to solve very complex tasks that once were thought to be out of the realm of machines. However, while humans and animals learn incrementally during their lifetimes and exploit their experience to solve new tasks, standard deep learning methods specialize to solve only one task at a time and whatever information they acquire is hardly reusable in new situations. Given that any artificial agent would need such a generalization ability to deal with the complexities of the world, it is critical to understand what mechanisms give rise to this ability. We argue that one of the mechanisms humans rely on is the use of discrete conceptual representations to encode their sensory inputs. These representations group similar inputs in such a way that combined they provide a level of abstraction that is transverse to a wide variety of tasks, filtering out irrelevant information for their solution. Here, we show that it is possible to learn such concept-like representations by self-supervision, following an information-bottleneck approach, and that these representations accelerate the transference of skills by providing a prior that guides the policy optimization process. Our method is able to learn useful concepts in locomotive tasks that significantly reduce the number of optimization steps required, opening a new path to endow artificial agents with generalization abilities.

Multi-label sentences (text) in the clinical domain result from the rich description of scenarios during patient care. The state-of-theart methods for assertion detection mostly address this task in the setting of a single assertion label per sentence (text). In addition, few rules based and deep learning methods perform negation/assertion scope detection on single-label text. It is a significant challenge extending these methods to address multi-label sentences without diminishing performance. Therefore, we developed a convolutional neural network (CNN) architecture to localize multiple labels and their scopes in a single stage end-to-end fashion, and demonstrate that our model performs atleast 12% better than the state-of-the-art on multi-label clinical text.

Their importance has been highlighted in recent times, when ICUs around the world have been overrun by the COVID-19 pandemic [1, 2]. It is in times like these when research into ways to adequately manage scarce critical care resources must be even more vigorously pursued, in order to offer additional tools that support medical decisions and allow for the effective benchmark of clinical practice. The issue of mortality prediction in the ICU has been approached from a statistical standpoint by means of risk prediction models like APACHE, SAPS, MODS, among others [3]. These models use a set of physiological predictors, demographic factors, and the occurrence of certain chronic conditions, to estimate a score that serves as a proxy for the likelihood of death of ICU patients. Because of the relatively straightforward way of interpreting results, simple statistical approaches such as logistic regression are the go-to modeling techniques used to estimate mortality probability and the importance of the predictors involved. On the other hand, the simplicity of the models also mean that their limited expressiveness may not accurately represent the possibly nonlinear dynamics of mortality prediction. Given this, high-capacity machine learning models might be useful to increase predictive performance.

Particularly in Natural Language Processing (NLP), the necessity of a good representation of the words that achieves some implicit context understanding is important [Turian et al., 2010]. A typical representation to feed in a machine learning model is the binary one-hot vector, in this case, an array with as many positions as words in the vocabulary is created, and the words are represented by a version of the array containing a one digit in the position corresponding to the word. For example, the sentence "I like eating bread and eating cheese", would have as vocabulary the set "I", "like", "eating", "bread", "and", "cheese", thus the representation of this words would be 6-dimensional binary one-hot vectors like "I" 100000, "like" 010000, cheese 000001. As you can imagine, this representation has no context understanding at all, since all words are completely independent, "bread" and "cheese" are as different between them as "I" and "like", which for a human is not the case.

Markov decision processes are of major interest in the planning community as well as in the model checking community. But in spite of the similarity in the considered formal models, the development of new techniques and methods happened largely independently in both communities. This work is intended as a beginning to unite the two research branches. We consider goal-reachability analysis as a common basis between both communities. The core of this paper is the translation from Jani, an overarching input language for quantitative model checkers, into the probabilistic planning domain definition language (PPDDL), and vice versa from PPDDL into Jani. These translations allow the creation of an overarching benchmark collection, including existing case studies from the model checking community, as well as benchmarks from the international probabilistic planning competitions (IPPC). We use this benchmark set as a basis for an extensive empirical comparison of various approaches from the model checking community, variants of value iteration, and MDP heuristic search algorithms developed by the AI planning community. On a per benchmark domain basis, techniques from one community can achieve state-ofthe-art performance in benchmarks of the other community. Across all benchmark domains of one community, the performance comparison is however in favor of the solvers and algorithms of that particular community. Reasons are the design of the benchmarks, as well as tool-related limitations. Our translation methods and benchmark collection foster crossfertilization between both communities, pointing out specific opportunities for widening the scope of solvers to different kinds of models, as well as for exchanging and adopting algorithms across communities.

A preliminary test in only eight volunteers suggests the first coronavirus vaccine to be tested in people seems to be safe and can stimulate an immune response against the virus. Antibodies generated by the volunteers were able to stop the virus from replicating in human cells in the laboratory and the levels of antibodies in their blood were similar to those previously detected in recovered covid-19 patients. Tal Zaks of Moderna, the US firm developing the vaccine, said that if the next stages go well, it could be widely available by the end of this year or early next year. The US stock market was up sharply today following the announcement. However, it remains to be seen if such speedy testing and manufacturing of a vaccine is really possible – no vaccine has ever been produced in less than five years. Meanwhile, a trial of another vaccine, developed by researchers at the University of Oxford found it wasn't able to stop six rhesus macaque monkeys from becoming infected with the ...

SparkCognition, a leading industrial artificial intelligence (AI) company, is pleased to announce that Japanese AI and Fintech company, MILIZE Co., Ltd. will offer Japanese financial institutions fraud detection and anti-money laundering solutions. These solutions will be built using the automated machine learning software of SparkCognition. With the enormous increase of online payment, internet banking, and QR code payments, illegal use of credit cards is on the rise. However, there are not many Japanese companies that have introduced advanced solutions for fraud detection that currently exist internationally. In addition, financial authorities and institutions around the world are expected to report strengthened measures against money laundering in August 2020. As a result, taking these steps against money laundering has become an urgent management issue in Japanese financial institutions.