In [Meilijson and Ruppin, 1993] we presented a methodological framework describing the two-iteration performance of Hopfield(cid:173) like attractor neural networks with history-dependent, Bayesian dynamics. We now extend this analysis in a number of directions: input patterns applied to small subsets of neurons, general con(cid:173) nectivity architectures and more efficient use of history. We show that the optimal signal (activation) function has a slanted sigmQidal shape, and provide an intuitive account of activation functions with a non-monotone shape.

Facebook claims that its new artificial intelligence can predict the way drugs interact with each other inside cells quicker than existing methods, enabling speedier discovery of new drug combinations to treat illnesses like cancer, but some researchers say it may not translate into results that will be useful in humans. The system, developed by Facebook AI Research and the Helmholtz Centre in Munich, Germany, is claimed to be the first easy-to-use AI model able to estimate how different drugs will work in the body. It could speed up our ability to uncover new treatments for diseases like cancer. "Drug research often takes half a decade to develop a compound," says Fabian Theis at the Helmholtz Centre, one of the authors of the work. The model works by measuring how individual cells change in response to treatment from a particular set of drugs and recording those responses.

Using artificial intelligence (AI), European researchers have developed an algorithm that they say successfully detects molecular changes in tumor cells and tissues from microscopic slides in many different cancers. "What is quite remarkable is that our algorithm can automatically link the histological appearance of almost any tumor with a very broad set of molecular characteristics and with patient survival," said Moritz Gerstung, PhD, group leader at EMBL European Bioformatics Institute. Institute researchers collaborated on the study with scientists from the Wellcome Sanger Institute and Addenbrooke's Hospital in Cambridge, UK. The pan-cancer analysis is believed to be the largest to date to train computer vision to "see" and combine digital pathology with the genetic changes that occur in cells as malignancies take hold. Ordinarily, histopathologists examine the appearance of cancer tissue under a microscope first, then geneticists perform molecular sequencing separately to analyze changes in the genetic code.

We implement and study a computational model of Stevens' [19921 theory of the pathogenesis of schizophrenia. This theory hypothesizes thatthe onset of schizophrenia is associated with reactive synaptic regeneration occurring in brain regions receiving degenerating temporallobe projections. Concentrating on one such area, the frontal cortex, we model a frontal module as an associative memory neural network whose input synapses represent incoming temporal projections. We analyze how, in the face of weakened external input projections, compensatory strengthening of internal synaptic connections and increased noise levels can maintain memory capacities(which are generally preserved in schizophrenia). However, These compensatory changes adversely lead to spontaneous, biasedretrieval of stored memories, which corresponds to the occurrence of schizophrenic delusions and hallucinations without anyapparent external trigger, and for their tendency to concentrate onjust few central themes. Our results explain why these symptoms tend to wane as schizophrenia progresses, and why delayed therapeuticalintervention leads to a much slower response.