The human visual system is well-tuned to detect faces of all shapes and sizes. While this brings obvious survival advantages, such as a better chance of spotting unknown predators in the bush, it also leads to spurious face detections. "Face pareidolia" describes the perception of face-like structure among otherwise random stimuli: seeing faces in coffee stains or clouds in the sky. In this paper, we study face pareidolia from a computer vision perspective. We present an image dataset of "Faces in Things", consisting of five thousand web images with humanannotated pareidolic faces. Using this dataset, we examine the extent to which a state-of-the-art human face detector exhibits pareidolia, and find a significant behavioral gap between humans and machines. We find that the evolutionary need for humans to detect animal faces, as well as human faces, may explain some of this gap. Finally, we propose a simple statistical model of pareidolia in images. Through studies on human subjects and our pareidolic face detectors we confirm a key prediction of our model regarding what image conditions are most likely to induce pareidolia.

We propose CHiLL (Crafting High-Level Latents), an approach for natural-language specification of features for linear models. CHiLL prompts LLMs with expert-crafted queries to generate interpretable features from health records. The resulting noisy labels are then used to train a simple linear classifier. Generating features based on queries to an LLM can empower physicians to use their domain expertise to craft features that are clinically meaningful for a downstream task of interest, without having to manually extract these from raw EHR. We are motivated by a real-world risk prediction task, but as a reproducible proxy, we use MIMIC-III and MIMIC-CXR data and standard predictive tasks (e.g., 30-day readmission) to evaluate this approach. We find that linear models using automatically extracted features are comparably performant to models using reference features, and provide greater interpretability than linear models using "Bag-of-Words" features. We verify that learned feature weights align well with clinical expectations.

In clinical procedures of angioplasty (i.e., open clogged coronary arteries), devices such as balloons and stents need to be placed and expanded in arteries under the guidance of X-ray fluoroscopy. Due to the limitation of X-ray dose, the resulting images are often noisy. To check the correct placement of these devices, typically multiple motion-compensated frames are averaged to enhance the view. Therefore, device tracking is a necessary procedure for this purpose. Even though angioplasty devices are designed to have radiopaque markers for the ease of tracking, current methods struggle to deliver satisfactory results due to the small marker size and complex scenes in angioplasty. In this paper, we propose an end-to-end deep learning framework for single stent tracking, which consists of three hierarchical modules: U-Net based landmark detection, ResNet based stent proposal and feature extraction, and graph convolutional neural network (GCN) based stent tracking that temporally aggregates both spatial information and appearance features. The experiments show that our method performs significantly better in detection compared with the state-of-the-art point-based tracking models. In addition, its fast inference speed satisfies clinical requirements.

Artificial intelligence, robotics and other new technologies are revolutionizing the field of cardiology, according to Rambam Health Care Campus Prof. Rafael Beyar. "Cardiology has been going through dramatic changes in the past few years," says Beyar. "There are several fields where we are witnessing those changes. First I would refer to digital health, and the development of better and quicker tools, based on artificial intelligence, for diagnosing heart attacks, strokes, pains and so on. "Nowadays, there are many ways to monitor patients, from smartwatches to patches to special cameras that can be installed in one's home, all of which collect information that can be sent to hospitals." Beyar is one of the organizers of the conference Innovation in Cardiovascular Interventions that will take place in Tel Aviv on December 5-7.

It is being used during Optical Coherence Tomography (OCT) catheterisation, which allows cardiologists to see inside the arteries of patients for more accurate placement of stents, which are used to treat blockages. When excess calcium accumulates in the blood and combines with cholesterol it forms plaque which adheres to the walls of arteries. These deposits can cause partial or complete blockage. OCT is routinely used to take images of the eyes of patients with glaucoma but is now increasingly being used to treat patients with heart disease. Dr Stuart Watkins and Dr Margaret McEntegart, consultant cardiologists at the Golden Jubilee Hospital, are now using the most advanced catheter on patients and a live recording of one of their latest cases will be used to educate cardiologists worldwide.

Colby College has hired a language processing expert to lead its newly formed Davis Institute for Artificial Intelligence. Amanda Stent, considered one of the country's leading authorities on natural language processing – which gives computers the ability to understand human text and spoken words – will start in October. Stent most recently served as the natural language processing architect at Bloomberg L.P., where she led the People and Language AI Team. Stent has authored or co-authored more than 100 papers on natural language processing and is a regular speaker on the subject. She was also involved in the CALO (Cognitive Assistant that Learns and Organizes) project that led to a range of AI applications, including the well-known virtual assistant Siri.

A new artificial intelligence technology that will help cardiologists to fit stents faster and with more precision will be used to treat heart attack patients across the country. Patients at the Royal Free hospital in London are the first in the country to be treated with the AI-driven keyhole procedure before it is introduced at 20 other sites. The pioneering technology will help cardiologists to make quicker and more accurate decisions while fitting a stent, a small tube of stainless steel mesh, for coronary artery disease, a procedure known as an angioplasty. It is one of many innovative technologies being used across the NHS to help get routine services back on track after the pandemic.

As a hybrid physician/engineer, I spend a lot of time pondering how new platforms can empower doctors. I am particularly excited about the potential of smart speakers coupled with advances in A.I. and natural language processing (also looking at you, blockchain). I am bullish on conversational agents in general, previously building an iOS chatbot powered by Watson that simulates a human radiologist. Chatbots are cool and useful, but voice -- that might be magic. Sensing potential, I decided to hunker down with my trusty corgi, drink a bunch of coffee, and start building the cool voice tools I want to use in my own clinical practice.

As more tools become available to create AI models, it has become easier for companies to harness the power of machine learning for their applications. What once required deep domain expertise to execute has been made easier by libraries and frameworks, such as Google's TensorFlow. To be clear, none of it is'easy,' but it may well be that the hardest part of the AI equation is acquiring, wrangling and, perhaps most poignantly, cleaning the data required to do the job. Engineers without experience in AI may well underestimate the time and effort required to get data to a point where AI will make the greatest impact, where the model will be as powerful and predictive as it can be. We talked to many data scientists and engineers who estimated that, on a given AI project, corralling, moving (these datasets can be unwieldy in their size), checking and organizing the data often comprises 70% to 80% of the time spent on a project.

That's the aim of a device that could help people control robotic limbs using thought alone – without the need for brain surgery. The device will be trialled in people with paralysis next year. Several groups are developing brain-machine interfaces that allow people who are paralysed to operate a bionic exoskeleton just by thinking about it. These devices decode electrical brain signals and translate them into movement of robotic limbs. Usually, brain signals are detected via electrodes attached to the scalp or implanted directly in the brain.