Transparency in artificial intelligence has become a question of provenance. Provenance tells you where a piece of content came from; it does not tell you why what it asserts was decided.
A number with no recorded reason may end up deciding for a patient Leer más »
Bold projections that artificial intelligence will accelerate scientific discovery have raced ahead of evidence from working scientists, and the field still lacks large-scale, scientist-in-the-loop tests of these claims. Here we mount the largest such evaluation to date and map what AI cannot yet do for science. We invited authors of 121,640 recent preprints across biology, medicine, chemistry, and the social sciences to judge ideas that large language models (LLMs) generated from the context and puzzles of their own papers. 6,749 scientists returned 25,139 sets of ratings on novelty, empirical feasibility, probability of being true, and favorability of adoption. Three patterns emerge. First, non-reasoning LLMs collapse into a narrow «hivemind» of similar ideas; reasoning models roam a wider hypothesis space, yet no model class spontaneously proposes null hypotheses — a move humans make more freely. Second, scientists reward ideas that resemble their own and prize probability over novelty, though social scientists tolerate risk more readily than life scientists. Senior social scientists are the harshest critics, and their skepticism is well-earned: LLMs falter most in pluralistic fields like the social sciences that demand context-aware interpretation and evolving theories. Third, automated evaluators on which the community currently relies — LLM-as-a-judge, artificial metrics, and even state-of-the-art (SOTA) models — agree only weakly with expert judgment, and retrieval augmentation and scientist persona prompting yield only marginal gains. A Qwen3-14B reward model we post-trained on human ratings captures field taste nuances, beats SOTA models by up to 27%, and closes the gap to the inter-rater consistency of independent peer reviewers. For all the hype, today’s scientific AI still represents a collaborator whose imagination, outputs and judgment benefit from human grounding.
Contemporary AI lacks the imagination to diverge or negate in science Leer más »
Large language models (LLMs) are designed to generate answers to user prompts, which often drives them to respond even when uncertainty is high, information is incomplete, or a refusal would be more appropriate. In healthcare, this tendency can be dangerous: confidently stated but inaccurate medical advice can cause significant harm, making the ability to abstain especially important. In this paper, we review studies investigating LLM abstention behaviors in healthcare. The literature highlights two main motivations: (1) uncertainty-driven abstention, where the model withholds a response when confidence is low, and (2) safety-driven abstention, where the model declines to provide potentially harmful information. Most existing mechanisms are extrinsic and rely on auxiliary tools to determine when to abstain. We find that state-of-the-art LLMs still struggle to refuse inappropriate prompts, and that few benchmarks evaluate abstention in realistic medical scenarios, where performance lags behind other domains. Building on these findings, we introduce a decision-theoretic formalization of abstention that models the trade-off between answering and withholding responses under uncertainty and potential harm. Based on this formulation, we present MedSAFE, a framework for evaluating abstention in clinical dialogs, and demonstrate its operationalization through a proof-of-concept pilot across clinical scenarios derived from the review.
Deep learning models that predict functional genomic measurements from DNA sequences are powerful tools for deciphering the genetic regulatory code. Existing methods involve a trade-off between input sequence length and prediction resolution, thereby limiting their modality scope and performance1,2,3,4,5. We present AlphaGenome, a unified DNA sequence model, which takes as input 1 Mb of DNA sequence and predicts thousands of functional genomic tracks up to single-base-pair resolution across diverse modalities. The modalities include gene expression, transcription initiation, chromatin accessibility, histone modifications, transcription factor binding, chromatin contact maps, splice site usage and splice junction coordinates and strength. Trained on human and mouse genomes, AlphaGenome matches or exceeds the strongest available external models in 25 of 26 evaluations of variant effect prediction. The ability of AlphaGenome to simultaneously score variant effects across all modalities accurately recapitulates the mechanisms of clinically relevant variants near the TAL1 oncogene6. To facilitate broader use, we provide tools for making genome track and variant effect predictions from sequence.
Advancing regulatory variant effect prediction with AlphaGenome Leer más »
Systems for diagnosis, prognosis and imaging have repeatedly been deployed or promoted on the strength of performance that proved fragile under independent testing. A model that performs well in the hospital where it was born has proven only one thing: that it works at home. Consider the most instructive failure in recent clinical AI. The
External validation is not a bureaucratic detail Leer más »
In 2019, a research team led by the physician and economist Ziad Obermeyer reverse-engineered a commercial algorithm already running quietly across the United States health system. The tool, sold by Optum, helped decide which patients — out of a population of roughly 200 million a year — would be flagged for extra medical attention. To
A medical algorithm must not confuse cost with health Leer más »
In an internal review that later became public, IBM’s Watson for Oncology was shown a patient much like thousands treated every day: a 65-year-old man with lung cancer who was also suffering from severe bleeding. The system recommended bevacizumab — a drug that carries an explicit warning against use in patients with severe hemorrhage, precisely
When AI sounds certain but should say “I do not know” Leer más »
In this presentation, the current issues to make federated learning flawlessly useful in the real world will be briefly overviewed. They are related to data/system heterogeneity, client management, traceability, and security. Also, we introduce the modularized federated learning framework, we currently develop, to experiment various techniques and protocols to find solutions for aforementioned issues. The framework will be open to public after development completes.
Federated Learning: Issues in Medical Application Leer más »
Underpinning the latest version of AlphaFold is a novel machine learning approach that incorporates physical and biological knowledge about protein structure, leveraging multi-sequence alignments, into the design of the deep learning algorithm.
Highly accurate protein structure prediction with AlphaFold Leer más »
Steganography is the science of hiding a secret message within an ordinary public message. Over the years, steganography has been used to encode a lower resolution image into a higher resolution image by simple methods like LSB manipulation. We aim to utilize deep neural networks for the encoding and decoding of multiple secret images inside a single cover image of the same resolution.
Multi-Image Steganography Using Deep Neural Networks Leer más »