Integrating high-level semantically correlated contents and low-level anatomical features is of central importance in medical image segmentation. Towards this end, recent deep learning-based medical segmentation methods have shown great promise in better modeling such information. However, convolution operators for medical segmentation typically operate on regular grids, which inherently blur the high-frequency regions, i.e., boundary regions. In this work, we propose MORSE, a generic implicit neural rendering framework designed at an anatomical level to assist learning in medical image segmentation. Our method is motivated by the fact that implicit neural representation has been shown to be more effective in fitting complex signals and solving computer graphics problems than discrete grid-based representation. The core of our approach is to formulate medical image segmentation as a rendering problem in an end-to-end manner. Specifically, we continuously align the coarse segmentation prediction with the ambiguous coordinate-based point representations and aggregate these features to adaptively refine the boundary region. To parallelly optimize multi-scale pixel-level features, we leverage the idea from Mixture-of-Expert (MoE) to design and train our MORSE with a stochastic gating mechanism. Our experiments demonstrate that MORSE can work well with different medical segmentation backbones, consistently achieving competitive performance improvements in both 2D and 3D supervised medical segmentation methods. We also theoretically analyze the superiority of MORSE.

Medical data often exhibits long-tail distributions with heavy class imbalance, which naturally leads to difficulty in classifying the minority classes (i.e., boundary regions or rare objects). Recent work has significantly improved semi-supervised medical image segmentation in long-tailed scenarios by equipping them with unsupervised contrastive criteria. However, it remains unclear how well they will perform in the labeled portion of data where class distribution is also highly imbalanced. In this work, we present ACTION++, an improved contrastive learning framework with adaptive anatomical contrast for semi-supervised medical segmentation. Specifically, we propose an adaptive supervised contrastive loss, where we first compute the optimal locations of class centers uniformly distributed on the embedding space (i.e., off-line), and then perform online contrastive matching training by encouraging different class features to adaptively match these distinct and uniformly distributed class centers. Moreover, we argue that blindly adopting a constant temperature $\tau$ in the contrastive loss on long-tailed medical data is not optimal, and propose to use a dynamic $\tau$ via a simple cosine schedule to yield better separation between majority and minority classes. Empirically, we evaluate ACTION++ on ACDC and LA benchmarks and show that it achieves state-of-the-art across two semi-supervised settings. Theoretically, we analyze the performance of adaptive anatomical contrast and confirm its superiority in label efficiency.

It uses three high-resolution cameras to peer down at the ground below. Lit by synchronized strobe lights, an onboard computer creates a digital image of each seedling as it glides by, comparing them with all the greenery it might reasonably find in a field of rich Salinas valley farmland two hours south of San Francisco. "It puts a dot on the stem and maps around it," says Todd Rinkenberger of FarmWise, the robot's maker. "Now it knows what's plant. Everything else is a weed."

The moment that Love Island fans have been waiting for is almost finally here, with Season 10 finally kicking off on Monday. To celebrate the imminent launch, a man has used AI to create the'perfect' Love Island contestants. Duncan Thomsen, 53, a freelance film editor from Brighton, trawled back through all previous nine series to see which couples have won the show so far. Then, using AI, he mashed photos of the winners together to make generic islanders, solely based on their looks. So, are his creations your type on paper?

An MRI scan showing the brain of a person with epilepsy.Credit: BSIP/Universal Images Group via Getty Virtual models representing the brains of people with epilepsy could help to enable more-effective treatments of the disease by showing neurosurgeons precisely which zones are responsible for seizures. The models, created using a computational system known as the Virtual Epileptic Patient (VEP), have been developed as part of the Human Brain Project (HBP), a 10-year European initiative focused on digital brain research. The approach is being tested in a clinical trial called EPINOV, to evaluate whether it improves the success rate of epilepsy surgeries. "It's an example of personalized medicine," says Aswin Chari, a neurosurgeon at University College London. VEP uses "the patient's own brain scans [and] the patient's own brainwave-recording data to build a model and improve our understanding of where their seizures are coming from".

Two years ago this weekend, GPT-3 was introduced to the world. You may not have heard of GPT-3, but there's a good chance you've read its work, used a website that runs its code, or even conversed with it through a chatbot or a character in a game. GPT-3 is an AI model -- a type of artificial intelligence -- and its applications have quietly trickled into our everyday lives over the past couple of years. In recent months, that trickle has picked up force: more and more applications are using AI like GPT-3, and these AI programs are producing greater amounts of data, from words, to images, to code. A lot of the time, this happens in the background; we don't see what the AI has done, or we can't tell if it's any good.

It's my first day "mindfiling," and I guess that's the sort of maturity you'd expect from a healthy 28-year-old considering his mortality. Mindfiling is a practice from the techno-religious faith movement Terasem, which celebrates personal cyberconsciousness. Its motto is "life is purposeful, death is optional, God is technological, love is essential." Mindfiling is a central daily act of uploading data about yourself to be stored until the resulting model of your mind and consciousness can be reconstructed and uploaded into an artificial body. It may be an act best understood in light of Ray Kurzweil's 2005 book The Singularity Is Near, in which he predicted A.I. would replicate and even outstrip human intelligence by the 2020s.

Cambrian-AI Research Sr. Analyst Gary Fritz contributed this blog. AI has been advancing dramatically in the last decade. It is able to solve classes of problems (facial recognition, machine translation, autonomous vehicles, and others) that were not suitably handled by traditional methods. But there are many important problems that cannot be addressed by traditional or AI approaches, either within a reasonable timeframe or, possibly, at all. Some of these problems can be attacked with Quantum Computing when that technology comes out of research and becomes more widely available.

The increasing availability of data presents an opportunity to calibrate unknown parameters which appear in complex models of phenomena in the biomedical, physical and social sciences. However, model complexity often leads to parameter-to-data maps which are expensive to evaluate and are only available through noisy approximations. This paper is concerned with the use of interacting particle systems for the solution of the resulting inverse problems for parameters. Of particular interest is the case where the available forward model evaluations are subject to rapid fluctuations, in parameter space, superimposed on the smoothly varying large scale parametric structure of interest. Multiscale analysis is used to study the behaviour of interacting particle system algorithms when such rapid fluctuations, which we refer to as noise, pollute the large scale parametric dependence of the parameter-to-data map. Ensemble Kalman methods (which are derivative-free) and Langevin-based methods (which use the derivative of the parameter-to-data map) are compared in this light. The ensemble Kalman methods are shown to behave favourably in the presence of noise in the parameter-to-data map, whereas Langevin methods are adversely affected. On the other hand, Langevin methods have the correct equilibrium distribution in the setting of noise-free forward models, whilst ensemble Kalman methods only provide an uncontrolled approximation, except in the linear case. Therefore a new class of algorithms, ensemble Gaussian process samplers, which combine the benefits of both ensemble Kalman and Langevin methods, are introduced and shown to perform favourably.

A smartphone that can warn you not to send a text while you're upset? Early in my career--back in the stone age before computers and smartphones--I worked in environments where memos were a primary means of communication. Sure, my colleagues and I could talk face-to-face, but the culture of the time was to memorialize much of our interaction in writing. Believe it or not, there were some advantages in what now seems such an archaic practice. Unlike texts and emails--where one tap of the "send" button can fill you with instant regret--the old-fashioned memo provided a cushion of safety, a chance to reconsider.