How The Pitt's AI Drama is Playing Out in Real Hospitals In Thursday's episode of The Pitt, the long-simmering tensions over the use of AI at the Pittsburgh Trauma Medical Center boiled over. In season two of the five-time Emmy winning medical drama, a new attending physician, Baran Al-Hashimi (Sepideh Moafi), is determined to improve efficiencies at the hospital. She tells her skeptical staff that new AI systems can cut down their time spent on charting by 80%, allowing them to spend more time both at the bedside and at home. But in episode six, doctors discover that the AI tool has made up false details about a patient and confused "urology" for "neurology." "AI's two percent error rate is still better than dictation," Al-Hashimi says, adding that it needs to be proofread for errors.

We contribute a novel approach to understand, dialogue, plan, and execute complex sentences to command a mobile service robot. We define a complex command as a natural language sentence consisting of sensing-based conditionals, conjunctions, and disjunctions. We introduce a flexible template-based algorithm to extract such structure from the parse tree of the sentence. As the complexity of the command increases, extracting the right structure using the template-based algorithm decreases becomes more problematic. We introduce two different dialogue approaches that enable the user to confirm or correct the extracted command structure. We present how the structure used to represent complex commands can be directly used for planning and execution by the service robot.

Data intensive solutions, such as solutions that include machine learning components, are becoming more and more prevalent. The standard way of developing such solutions is to train machine learning models with manually annotated or labeled data for a given task. This methodology assumes the existence of ample human annotated data. Unfortunately, this is often not the case, due to imbalanced distribution of classes and lack of human annotation resources. This challenge is exasperated when thousands of hierarchical classes are introduced.

CoBot is a service mobile robot that has been continuously deployed for extended periods of time in a multi-floor office-style building. While moving in the building CoBot, is able to perform multiple tasks for its users; the robot is able to autonomously navigate to any of the rooms in the building, to deliver objects and messages and to escort visitors to their destination. While apparently very different, all the tasks CoBot is able to perform require the robot to move from one location to another. In our effort to make CoBot increasingly available to users in the building, we recently enabled it to understand spoken commands, an approach for Learning and using task-relevant Knowledge from human-robot Dialog and access to the Web.

Inherently, the legal domain contains a vast amount of data in text format. Therefore it requires the application of Natural Language Processing (NLP) to cater to the analytically demanding needs of the domain. The advancement of NLP is spreading through various domains, such as the legal domain, in forms of practical applications and academic research. Identifying critical sentences, facts and arguments in a legal case is a tedious task for legal professionals. In this research we explore the usage of sentence embeddings for multi-class classification to identify critical sentences in a legal case, in the perspective of the main parties present in the case. In addition, a task-specific loss function is defined in order to improve the accuracy restricted by the straightforward use of categorical cross entropy loss.

Ronen Lavi and Shay Perera have spent years working to develop and deploy AI in one of the most demanding environments in the world--the elite intelligence units of the Israel Defense Forces (IDF). Lavi established and led the AI Lab of Israel's Military Intelligence and Perera served there as manager of machine learning and computer vision research and development. After being awarded a National Security Award in 2018, they left the IDF to launch a startup, as many Israelis with similar experience and skills have done before them. The rapid digital transformation of the healthcare industry worldwide, the proliferation of healthcare data, the increasing complexity of healthcare (including its administration), the dearth of qualified personnel--and the Covid pandemic--have all contributed to a rising demand for AI solutions, intended to assist with detection, diagnosis, treatment, preventive care and wellness. The wealth of data that is produced by digitized medical records is what modern AI approaches (deep learning) require so they can "learn" from examples, automate certain decisions, and provide a helping hand to physicians and healthcare staff.

ABSTRACT Transient phenomena associated with forward biased silicon p - n - n structures at 4.2K show remarkable similarities with biological neurons. The devices play a role similar to the two-terminal switching elements in Hodgkin-Huxley equivalent circuit diagrams. The devices provide simpler and more realistic neuron emulation than transistors or op-amps. They have such low power and current requirements that they could be used in massive neural networks. Some observed properties of simple circuits containing the devices include action potentials, refractory periods, threshold behavior, excitation, inhibition, summation over synaptic inputs, synaptic weights, temporal integration, memory, network connectivity modification based on experience, pacemaker activity, firing thresholds, coupling to sensors with graded signal outputs and the dependence of firing rate on input current. Transfer functions for simple artificial neurons with spiketrain inputs and spiketrain outputs have been measured and correlated with input coupling.

ABSTRACT Transient phenomena associated with forward biased silicon p - n - n structures at 4.2K show remarkable similarities with biological neurons. The devices play a role similar to the two-terminal switching elements in Hodgkin-Huxley equivalent circuit diagrams. The devices provide simpler and more realistic neuron emulation than transistors or op-amps. They have such low power and current requirements that they could be used in massive neural networks. Some observed properties of simple circuits containing the devices include action potentials, refractory periods, threshold behavior, excitation, inhibition, summation over synaptic inputs, synaptic weights, temporal integration, memory, network connectivity modification based on experience, pacemaker activity, firing thresholds, coupling to sensors with graded signal outputs and the dependence of firing rate on input current. Transfer functions for simple artificial neurons with spiketrain inputs and spiketrain outputs have been measured and correlated with input coupling.

ABSTRACT Transient phenomena associated with forward biased silicon p - n - n structures at 4.2K show remarkable similarities with biological neurons. The devices play a role similar to the two-terminal switching elements in Hodgkin-Huxley equivalent circuit diagrams. The devices provide simpler and more realistic neuron emulation than transistors or op-amps. They have such low power and current requirements that they could be used in massive neural networks. Some observed properties of simple circuits containing the devices include action potentials, refractory periods, threshold behavior, excitation, inhibition, summation over synaptic inputs, synaptic weights, temporal integration, memory, network connectivity modification based on experience, pacemaker activity, firing thresholds, coupling to sensors with graded signal outputsand the dependence of firing rate on input current. Transfer functions for simple artificial neurons with spiketrain inputs and spiketrain outputs have been measured and correlated with input coupling.