Using large language models (LLMs), computers are able to generate a written text in response to a us er request. As this pervasive technology can be applied in numerous contexts, this study analyses the written style of one LLM called GPT by comparing its generated speeches with those of the recent US presidents. To achieve this objective, the State of the Union (SOTU) addresses written by Reagan to Biden are contrasted to those produced by both GPT-3.5 and GPT-4.o versions. Compared to US presidents, GPT tends to overuse the lemma "we" and produce shorter messages with, on average, longer sentences. Moreover, GPT opts for an optimistic tone, opting more often for political (e.g., president, Congress), symbolic (e.g., freedom), and abstract terms (e.g., freedom). Even when imposing an author's style to GPT, the resulting speech remains distinct from addresses written by the target author. Finally, the two GPT versions present distinct characteristics, but both appear overall dissimilar to true presidential messages.

Recently several large language models (LLMs) have demonstrated their capability to generate a message in response to a user request. Such scientific breakthroughs promote new perspectives but also some fears. The main focus of this study is to analyze the written style of one LLM called ChatGPT 3.5 by comparing its generated messages with those of the recent US presidents. To achieve this objective, we compare the State of the Union addresses written by Reagan to Obama with those automatically produced by ChatGPT. We found that ChatGPT tends to overuse the lemma "we" as well as nouns and commas. On the other hand, the generated speeches employ less verbs and include, in mean, longer sentences. Even when imposing a given style to ChatGPT, the resulting speech remains distinct from messages written by the target author. Moreover, ChatGPT opts for a neutral tone with mainly positive emotional expressions and symbolic terms (e.g., freedom, nation). Finally, we show that the GPT's style exposes distinct features compared to real presidential addresses.

Different entities with the same name can be difficult to distinguish. Handling confusing entity mentions is a crucial skill for language models (LMs). For example, given the question "Where was Michael Jordan educated?" and a set of documents discussing different people named Michael Jordan, can LMs distinguish entity mentions to generate a cohesive answer to the question? To test this ability, we introduce a new benchmark, AmbigDocs. By leveraging Wikipedia's disambiguation pages, we identify a set of documents, belonging to different entities who share an ambiguous name. From these documents, we generate questions containing an ambiguous name and their corresponding sets of answers. Our analysis reveals that current state-of-the-art models often yield ambiguous answers or incorrectly merge information belonging to different entities. We establish an ontology categorizing four types of incomplete answers and automatic evaluation metrics to identify such categories. We lay the foundation for future work on reasoning across multiple documents with ambiguous entities.

While enabling accelerated acquisition and improved reconstruction accuracy, current deep MRI reconstruction networks are typically supervised, require fully sampled data, and are limited to Cartesian sampling patterns. These factors limit their practical adoption as fully-sampled MRI is prohibitively time-consuming to acquire clinically. Further, non-Cartesian sampling patterns are particularly desirable as they are more amenable to acceleration and show improved motion robustness. To this end, we present a fully self-supervised approach for accelerated non-Cartesian MRI reconstruction which leverages self-supervision in both k-space and image domains. In training, the undersampled data are split into disjoint k-space domain partitions. For the k-space self-supervision, we train a network to reconstruct the input undersampled data from both the disjoint partitions and from itself. For the image-level self-supervision, we enforce appearance consistency obtained from the original undersampled data and the two partitions. Experimental results on our simulated multi-coil non-Cartesian MRI dataset demonstrate that DDSS can generate high-quality reconstruction that approaches the accuracy of the fully supervised reconstruction, outperforming previous baseline methods. Finally, DDSS is shown to scale to highly challenging real-world clinical MRI reconstruction acquired on a portable low-field (0.064 T) MRI scanner with no data available for supervised training while demonstrating improved image quality as compared to traditional reconstruction, as determined by a radiologist study.

Multi-contrast MRI (MC-MRI) captures multiple complementary imaging modalities to aid in radiological decision-making. Given the need for lowering the time cost of multiple acquisitions, current deep accelerated MRI reconstruction networks focus on exploiting the redundancy between multiple contrasts. However, existing works are largely supervised with paired data and/or prohibitively expensive fully-sampled MRI sequences. Further, reconstruction networks typically rely on convolutional architectures which are limited in their capacity to model long-range interactions and may lead to suboptimal recovery of fine anatomical detail. To these ends, we present a dual-domain self-supervised transformer (DSFormer) for accelerated MC-MRI reconstruction. DSFormer develops a deep conditional cascade transformer (DCCT) consisting of several cascaded Swin transformer reconstruction networks (SwinRN) trained under two deep conditioning strategies to enable MC-MRI information sharing. We further present a dual-domain (image and k-space) self-supervised learning strategy for DCCT to alleviate the costs of acquiring fully sampled training data. DSFormer generates high-fidelity reconstructions which experimentally outperform current fully-supervised baselines. Moreover, we find that DSFormer achieves nearly the same performance when trained either with full supervision or with our proposed dual-domain self-supervision.

Over the last decade, though, the tech industry has downsized ultrasound scanners into devices resembling TV remotes. It has also created digital stethoscopes that can be paired with smartphones to create moving pictures and readouts. Proponents say these devices are nearly as easy to use as stethoscopes and allow doctors to watch the body in motion and actually see things such as leaky valves. "There's no reason you would listen to sounds when you can see everything," Topol said. At many medical schools, it's the newer devices that really get students' hearts pumping.

GUILFORD, Conn., Sept. 17, 2019 -- InveniAI Corporation, a global leader pioneering the application of artificial intelligence (AI) and machine learning (ML) to transform innovation across healthcare and other industries, today announced the expansion of its flagship AI- and ML-driven innovation monitoring platform, AlphaMeld . This expansion will provide a suite of products to facilitate the democratization of data analysis across the life sciences value chain and tap into InveniAI's extensive domain expertise and curated data sets that have been aligned with industry specific algorithms. CIMeld, an AlphaMeld powered product, will be launched at the upcoming PharmaCI USA 2019 Conference and Exhibition set for September 18-19, where competitive intelligence executives in pharma, biotech and medical devices meet at one of the largest CI assemblies in the world. CIMeld is a unique product that takes advantage of the digital data deluge and leverages AI-powered technology to triangulate signals from multiple sources to generate insights and identify opportunities and threats across diseases, companies, drug portfolios, and early innovation. The AlphaMeld suite of products is accessible via a secure, customized, cloud-based interface with the option to personalize the definition of success and failure that lends a unique competitive advantage to its users.

Artificial intelligence and machine learning are quickly becoming an integral part of healthcare delivery. Both on the clinical care and operational side of healthcare organizations, AI has is powering technology that keeps patients safe and improves efficiency for the revenue cycle, supply chain and more. Here are 100-plus companies in the healthcare space using artificial intelligence. To add a company to this list, contact Laura Dyrda at ldyrda@beckershealthcare.com. AiCure is an AI and advanced data analytics company that uses video, audio and behavioral data to better understand the connection between patients, disease and treatment. It allows physicians to have access to clinical and patient insights. Aiva Health developed Aiva, the first voice-powered care assistant.

"What I love," says Alyssa Siefert, Engineering Director at Yale Center for Biomedical Innovation and Technology (CBIT), "is a democratization of problem solving." Siefert is one of the lead organizers of the Yale Healthcare Hackathon, an event in its fifth year that brings together a diverse group of clinicians, engineers, designers, patients and community members Jan 19-21 at Yale School of Medicine to come up with solutions to healthcare challenges. Last year, the event had representatives from eight countries and two dozen universities, and those numbers have been on the rise. About half the participants are non-Yale. The main sponsor of this year's event is 4Catalyzer, a Guilford, Connecticut-based accelerator founded by Dr. Jonathan Rothberg, who serves as its Chief Strategy Officer, for launching new biomedical startups with a heavy emphasis on medical devices, artificial intelligence and big data.