Former US Attorney General Pam Bondi, who was removed from her role last month, has been diagnosed with thyroid cancer, according to multiple US outlets. Her diagnosis came shortly after President Donald Trump ousted her from the post of America's top law enforcement officer, according to Axios, which first reported the news of her illness. Bondi, 60, told CNN she is undergoing treatment and is still recovering from surgery that took place a few weeks ago, but is doing well. She is continuing to work despite the diagnosis, and will be joining the White House's new advisory council on AI, the Presidential Council of Advisors on Science and Technology. Podcast host and former White House adviser Katie Miller posted on social media that Pam has been quietly kicking cancer's ass the last few weeks, adding that Bondi has a heart of gold.

In multi-modal biomedical research, integrating high-dimensional genomic data with clinical baselines is essential for precision medicine. However, standard deep neural network approaches often entangle these modalities, obscuring the specific predictive impact of genetic features and leading to possibly suboptimal predictive performance. Motivated by the landmark METABRIC cohort primary breast tumors study, we propose the Stein-Encoder, a white-box supervised framework designed to isolate the genetic signal driving clinical outcomes conditional on nuisance covariates. By leveraging Stein's method and residualization techniques, our approach constructs an interpretable single index that summarizes relevant biological heterogeneity while flexibly incorporating clinical factors and can be used to improve downstream prediction. We establish theoretical guarantees for identification, consistency and efficiency improvement. Applied to the METABRIC cohort, the Stein-Encoder outperforms unsupervised benchmarks in predictive accuracy. Crucially, it achieves structural disentanglement by revealing response-specific biological mechanisms: we find that tumor size is driven primarily by mitotic networks, whereas prognostic indices rely on a distinct proliferation-versus-immune axis. This work contributes a unified, computationally efficient framework that bridges statistical rigor with the representational power of neural networks, enabling interpretable, task-specific and efficient compression of multi-modal health data for a wide range of precision medicine applications, beyond biomarker discovery.

Uta Frith seems remarkably cheerful and content for someone who's spent six decades trying and failing to get to grips with her life's obsession. "Very little has stood the test of time," she tells me as we sit down in her living room in a leafy estate in Harrow-on-the-Hill, London. Around us, high-ceilinged walls papered in a luxurious red print are barely visible between rammed bookshelves, several model brains and a collection of abstract art. Frith has been searching for the mechanisms that underpin the enigmatic condition of autism ever since she first met profoundly autistic children in the late 1960s. "We could identify them intuitively, but not really scientifically - and I have to say that this is, unfortunately, still the case." Still, Frith's influence on our ever-shifting understanding of autism has been monumental.

The Next Alzheimer's Breakthrough Will Take More Than Just Science At WIRED Health, pioneering Alzheimer's researcher John Hardy outlined the stakes--and next steps--of where treatment is headed next. Alzheimer's research is entering a new phase, as treatments that have taken decades to develop begin to reach patients . But getting those advances to people will depend on more than scientific progress alone, according to pioneering Alzheimer's researcher John Hardy . Speaking at WIRED Health in April, Hardy, chair of the Molecular Biology of Neurological Disease at University College London, said that alongside more effective drugs, better diagnosis and political will were still needed to improve treatment of Alzheimer's disease. "We've got to get better," he said.

Hyponatremia: Predict whether a hyponatremia lab comes back as normal (>=135 mmol/L), mild (>=130 and <135 mmol/L), moderate (>=125 and <130 mmol/L), or severe (<125 mmol/L). We consider all lab results coded as LOINC/LG11363-5, LOINC/2951-2, or LOINC/2947-0. Anemia: Predict whether an anemia lab comes back as normal (>=120 g/L), mild (>=110 and <120 g/L), moderate (>=70 and <110 g/L), or severe (<70 g/L). We consider all lab results coded as LOINC/LP392452-1. Please note that for the results of our baseline experiments in Section 5, we reframe these lab value tasks as binary classification tasks, where a label is "negative" if the result is normal and "positive" otherwise.

This paper proposes REFUEL, a reinforcement learning method with two techniques: reward shapingand feature rebuilding, to improve the performance of online symptom checking for disease diagnosis.

Intelligent education stands as a prominent application of machine learning. Within this domain, cognitive diagnosis (CD) is a key research focus that aims to diagnose students' proficiency levels in specific knowledge concepts. As a crucial task within the field of education, cognitive diagnosis encompasses two fundamental requirements: accuracy and fairness. Existing studies have achieved significant success by primarily utilizing observed historical logs of student-exercise interactions. However, real-world scenarios often present a challenge, where a substantial number of students engage with a limited number of exercises.