This is a month to look out for some powerful new books, with authors taking on challenges of all sorts and imagining whole new worlds. There are fresh ways to think about a cancer diagnosis, a book tackling the real inner world of hormones, in which we are all hormonal all the time, plus a major re-envisioning of the natural world where we abandon the shallows of competition for the depth and intricacies of connection and togetherness. It's quite hard going to get an up-to-date grip on human evolution, even for the best-briefed adult, so a book with sophisticated text and excellent illustrations and diagrams can only be a good thing. Especially if it is curated and edited by Alice Roberts, biological anthropologist, palaeopathologist, broadcaster - and professor of public engagement in science at the University of Birmingham, UK. She worked with a generous-sized international team of experts in many fields of human evolution, including archaeology, palaeontology, anthropology and cognitive science.

Conditional density estimation (CDE) - recovering the full conditional distribution of a response given tabular covariates - is essential in settings with heteroscedasticity, multimodality, or asymmetric uncertainty. Recent tabular foundation models, such as TabPFN and TabICL, naturally produce predictive distributions, but their effectiveness as general-purpose CDE methods has not been systematically evaluated, unlike their performance for point prediction, which is well studied. We benchmark three tabular foundation model variants against a diverse set of parametric, tree-based, and neural CDE baselines on 39 real-world datasets, across training sizes from 50 to 20,000, using six metrics covering density accuracy, calibration, and computation time. Across all sample sizes, foundation models achieve the best CDE loss, log-likelihood, and CRPS on the large majority of datasets tested. Calibration is competitive at small sample sizes but, for some metrics and datasets, lags behind task-specific neural baselines at larger sample sizes, suggesting that post-hoc recalibration may be a valuable complement. In a photometric redshift case study using SDSS DR18, TabPFN exposed to 50,000 training galaxies outperforms all baselines trained on the full 500,000-galaxy dataset. Taken together, these results establish tabular foundation models as strong off-the-shelf conditional density estimators.

Major leap towards reanimation after death as mammal's brain preserved A pig's brain has been frozen with its cellular activity locked in place and minimal damage. Could our brains one day be preserved in a way that locks in our thoughts, feelings and perceptions? An entire mammalian brain has been successfully preserved using a technique that will now be offered to people who are terminally ill. The intention is to preserve all the neural information thought necessary to one day reconstruct the mind of the person it once belonged to. "They would need to donate their brain and body for scientific research," says Borys Wróbel at Nectome in San Francisco, California, a research company focused on memory preservation.

Neural Information Processing Systems http://nips.cc/

Black-box optimisation is one of the important areas in optimisation.

The teacher-student framework, prevalent in semi-supervised semantic segmentation, mainly employs the exponential moving average (EMA) to update a single teacher's weights based on the student's. However, EMA updates raise a problem in that the weights of the teacher and student are getting coupled, causing a potential performance bottleneck. Furthermore, this problem may become more severe when training with more complicated labels such as segmentation masks but with few annotated data. This paper introduces Dual Teacher, a simple yet effective approach that employs dual temporary teachers aiming to alleviate the coupling problem for the student.

Toexploitthephenomenon and yet dismiss non-anomalous descent, SARAD performs anomaly detection via autoencoding in the association space.