Headlines

The FDA's TEMPO pilot now lists four devices, including an AI voice agent delivering therapy to Medicare patients without premarket authorization.

Artificial intelligence has transformed biomedical research over the past decade, but most AI systems in biology remain confined to narrow tasks, single data types, or one biological scale at a time. A new Perspective article published in Advanced Science proposes a radical conceptual remedy: a hypothetical “Full-Body AI Agent,” a supervisory multi-agent framework designed to […]

The durability of Meloni’s coalition has been a rarity in postwar Italy, where governments have often fallen before the end of their mandates.

The human gut is home to trillions of viruses, a sprawling community known as the virome, and yet it remains one of the least understood layers of the microbiome. A new review published in the journal Gut Pathogens argues that this viral darkness deserves far more clinical attention, particularly in patients undergoing hematopoietic stem cell […]

In a finding that could reshape how scientists think about repairing the damaged human heart, researchers at the University of Guelph have shown that the leopard gecko, a small terrestrial reptile best known as a popular pet, can regenerate its heart tissue after serious injury. Using a powerful technique known as quantitative proteomics, the team […]

In a development that could reshape how scientists probe the molecular machinery of legumes, researchers in Poland have created the first transgene-free knockout lines of a key transporter gene in the model legume Medicago truncatula, using a two-stage CRISPR/Cas9 strategy that promises to dramatically accelerate functional genomics in one of agriculture’s most important plant families. […]

The immune system is made up of specialized cells and proteins that help defend the body from pathogens and cancer. Most of the time, the immune system performs its job remarkably well. However, in some situations, the immune system doesn't work well enough or works in overdrive, causing disease.

A Cleveland Clinic first-in-human clinical trial has shown that a one-time infusion of a gene-editing therapy using CRISPR-Cas9 was effective and safe in reducing LDL ("bad") cholesterol and triglycerides in people with medication-resistant lipid disorders through one year of follow-up across all doses.

A lab-grown micro bladder may explain why urinary tract infections (UTIs) return so frequently, according to a study led by researchers at UCL, the University of Oxford and the University of Leicester.

As the human brain takes shape before birth, a remarkable type of stem cell helps give it many of the features that make it distinctly human.

Federal and state regulators are drawing new boundaries around medical AI, from how the tools are vetted to who bears responsibility when they fail.

The latest news and headlines from Yahoo! News. Get breaking news stories and in-depth coverage with videos and photos.

This is far from the only study to indicate that the children of long-lived individuals tend also to be long-lived. The open question is the degree to which this is genetic versus cultural transmission. The construction of very large population databases over the past few decades has produced evidence for genetics to play only a small role in life expectancy for the vast majority of people. Similarly the weight of evidence for lifestyle choice, particularly physical fitness, to be the presently dominant influence on life expectancy is compelling. Against that background, the hypothesis that centenarians are centenarians because they have favorable gene variants that might form the basis for age-slowing therapies faces an uphill battle. Centenarian lifespan extension is frequently accompanied by delayed onset of […]

Unauthorized stem cell treatment may expose patients to medical and financial risks, reinforcing the role of physicians in counseling those considering such treatment abroad.

For more than a century, the adult human heart has been viewed as a terminally differentiated organ, incapable of meaningful self-repair. Cardiomyocytes, the contractile cells that make up the bulk of cardiac tissue, exit the cell cycle shortly after birth and never meaningfully return. When a heart attack destroys a section of muscle, the organ […]

Pig kidneys successfully transplanted into humans  Financial TimesPatient Who Received Pig Kidney Becomes First To Progress to Human Kidney Transplant  Harvard Medical SchoolUse pig organs while patients wait for donors, say scientists  The TelegraphPig kidney kept a man off dialysis for nine months in world first  New ScientistPig kidney works in human body for 271 days - is this the future of organ transplants?  BBC

Pig kidneys successfully transplanted into humans  Financial TimesPatient Who Received Pig Kidney Becomes First To Progress to Human Kidney Transplant  Harvard Medical SchoolUse pig organs while patients wait for donors, say scientists  The TelegraphPig kidney kept a man off dialysis for nine months in world first  New ScientistPig kidney works in human body for 271 days - is this the future of organ transplants?  BBC

arXiv:2609.03865v1 Announce Type: new Abstract: Deng et al. discovered that Drt3b synthesizes protein-primed poly(AC) DNA without a nucleic-acid template. The structures reveal a finite protein architecture and product-associated contacts, but not the renewable dynamical rule that converts a bounded pocket into a long period-2 sequence. We propose that Drt3b is a regenerating protein-substrate free-energy register. The framework specifies a sampled quantum mechanics/molecular mechanics (QM/MM) potential of mean force and a preregistered, within-domain first-order electronic-descriptor map for edge-specific activation barriers; physical curvature is admitted only through a training-frozen Taylor-remainder bound that widens uncertainty but never corrects a held-out mean. Conditional hazards derived from those barriers are averaged over unresolved conformational, hydration, protonation, and metal-coordination trajectories to yield a semi-Markov event kernel. Pointwise ratios of competing

Hydrogels have long been considered one of the most promising material platforms for implantable bioelectronics. Soft, water-rich, and chemically similar to living tissue, they promise a gentle interface between electronics and biology. Yet despite decades of enthusiasm, hydrogel-based implants have remained largely confined to laboratory demonstrations, largely because of a single, stubborn failure mode: swelling. […]

Pig kidney works in human body for 271 days - is this the future of organ transplants?  BBCGenetically modified pig kidney keeps man alive for record nine months  The GuardianA Transplanted Pig Kidney Is Still Working After a Record-Setting 9 Months in a Patient  WIREDRecipients of gene-edited pig kidneys successfully receive human organs  STATIn a First, a Pig Kidney Kept a Patient Alive Until He Got a Human One  The New York Times

Pig kidney works in human body for 271 days - is this the future of organ transplants?  BBCGenetically modified pig kidney keeps man alive for record nine months  The GuardianA Transplanted Pig Kidney Is Still Working After a Record-Setting 9 Months in a Patient  WIREDRecipients of gene-edited pig kidneys successfully receive human organs  STATIn a First, a Pig Kidney Kept a Patient Alive Until He Got a Human One  The New York Times

Alexei Oreskovic / Fortune: AI scientist and Google researcher Ray Kurzweil joins BCI startup Subsense, which develops non-surgical BCI tech involving inhaled nanoparticles, as an advisor  —  Ray Kurzweil, the pioneering AI scientist, Google researcher, and “AI visionary,” is joining a Silicon Valley startup developing …

Fake doctor reviews are longer, more detailed, and more trusted than real ones, a gap generative AI is now making easier to exploit.

WashU Medicine surgeons report the first robotic combined liver-kidney transplant, replacing a 16-inch incision with one of about three inches.

Gene-edited pig kidneys could offer a lifeline to patients stuck waiting for a human donor.

In January 2026, Tim Andrews, who had received a genetically edited pig kidney in January 2025, became the recipient of a human kidney. A new Mass General Brigham paper, published in The Lancet, details Andrews' experience and how xenotransplantation could bridge the gap between end-stage kidney failure and a human transplant, allowing a patient to forgo dialysis in the meantime.

As the human brain takes shape before birth, a remarkable type of stem cell helps give it many of the features that make it distinctly human. These cells, called radial glia, make billions of decisions that determine which kinds of brain cells are created, when they appear and how the cortex—the seat of thought, memory and language—takes form.

Researchers have demonstrated a series of biological tests that could serve as "proxies" for measuring how different regenerative agriculture practices could improve soil carbon—years before it would be possible to measure the carbon directly. The technique offers new insights into how effective various regenerative practices are at improving soil health.

Artificial intelligence (AI) tools are overwhelmingly being used in mental health practices to support administrative tasks, but University of Queensland researchers have found they potentially lack privacy, accuracy and governance safeguards.

FDA warned an Arizona firm over unapproved umbilical cord products marketed for autism and ALS, citing unvalidated sterility processes.

A new risk stratification model developed by Northwestern Medicine scientists may help identify patients at lower risk of kidney transplant rejection, informing more personalized treatment and care strategies with fewer adverse side effects, as detailed in a recent study published in Nature Communications.

Senescent cells accumulate with age in tissues throughout the body, disrupting tissue structure and function with inflammatory secretions. The growing burden of senescent cells is an important contributing cause of degenerative aging. A cell becomes senescent in response to damage induced by some form of stress, or more often when a somatic cell reaches the Hayflick limit on cellular replication. In youth senescent cells are efficiently cleared by the immune system, but immune system aging allows senescent cells to linger and grow in number. Immune cells themselves are prone to reaching the Hayflick limit and entering replicative senescence because they respond to a variety of issues with greater replication to multiply their efforts. Research into neurodegenerative conditions and the state of the aging brain has […]

This eBook brings together expert perspectives on several of the most pressing analytical topics facing today’s cell and gene therapy developers. The post The Analytical Blueprint for Cell & Gene Therapy Success appeared first on GEN - Genetic Engineering and Biotechnology News.

The executive chair of Oracle and its CEO will be asked to justify why a project budgeted at $10bn is costing taxpayers $17bn more

Gene editing was supposed to be clean. The patent fight wasnt. For fourteen years, CRISPR has moved from a breakthrough scientific idea to a platform technology that promises cures, not just therapies. Investors backed the platform. Patients waited f... Source

In a rare case that is drawing attention across the respiratory medicine community, clinicians have documented the recurrence of pulmonary alveolar proteinosis more than three years after a patient received a bilateral lung transplant—a finding that challenges assumptions about what transplantation can and cannot fix, and that offers a roadmap for diagnosing a disease that […]

Metformin, a widely prescribed, safe and inexpensive drug for type 2 diabetes, is a first-line treatment that lowers blood sugar and improves the body's response to insulin. The drug appears to trick cells into sensing an energy shortage, similar to what happens during fasting or dieting. This activates a key cellular energy sensor called AMPK, which helps trigger cellular waste cleanup and supports cellular repair. Now, a recent review suggests metformin might also help counter some effects of aging.

uniQure filed for FDA approval of AMT-130, a one-time Huntington's gene therapy, and requested priority review. Here is what it means.

Long-lived bats are extreme outliers in the normal mammalian relationships between species body size, metabolic rate, and life span. They are small, have a high metabolic rate, and unlike near all other mammals with those characteristics, many bat species are long-lived. But bat species do exhibit a very wide range of life spans; near neighbor species can have very different paces of aging. Researchers here sequence the genomes of a number of different bat species in search of insight into the mechanisms driving longevity in bats. As is usually the case, the distant end goal of this sort of comparative biology of aging research is to find potential approaches to the development of longevity therapies. That is a long road, and little progress has been […]

arXiv:2609.01653v1 Announce Type: new Abstract: Diagnostic error carries a burden, while unconstrained large language models (LLMs) remain vulnerable to hallucination and weak integration of quantitative laboratory dynamics. We developed a decision-support pipeline combining disease-specific biomarker correlation graphs, ordinary differential equations (ODEs), deep sequence classification, and retrieval-augmented generation (RAG). For 103 disease classes from a full blood count (FBC) repository, biomarker networks were used as coupling matrices to generate 30 trajectories per disease (3,090 total). A one-dimensional convolutional neural network (CNN) and long short-term memory (LSTM) network classified disease trajectories and six dynamical clusters. A constrained GPT-4o-mini RAG layer used a 19-pattern BMJ Best Practice/NICE corpus to generate differential diagnoses evaluated for diagnostic suitability, evidential grounding, and clinical plausibility. Across five random-seed runs,

arXiv:2609.02113v1 Announce Type: cross Abstract: Classical and current quantum approaches to protein structure prediction (QPSP) face limitations, notably massive qubit requirements restricting near-term models to simplistic on-lattice simulations. We propose a logarithmic-scale variational quantum eigensolver (VQE) that reduces qubit requirements for N torsional degrees of freedom to O(log2N), enabling off-lattice, all-atom simulations. Our architecture extracts molecular torsions from relative phases in statevector simulations. On quantum hardware, a decoder maps the empirical cumulative distribution function (CDF) from basis-state probabilities to bounded torsional variables. These feed a classical algorithm to build heavy-atom coordinates. We use an EfficientSU2 ansatz and multi-stage relaxation to mitigate barren plateaus. Structures are evaluated via a custom hybrid quantum-classical Hamiltonian, alongside Rosetta and OpenMM benchmarks. Evaluation on chignolin and Trp-cage

arXiv:2609.02436v1 Announce Type: new Abstract: Collaborative Brain-Computer Interfaces (cBCIs) offer a promising mechanism to augment team decision-making, but existing approaches rely exclusively on evidence available only after a decision has been made and reported, such as reaction time or stated confidence. This limits their use to explaining or discounting a decision after the fact, rather than informing a team's response before it is finalised. We tested whether spatial-covariance EEG features could instead provide a genuinely pre-emptive signal of an operator's decision correctness, available within the response window itself, and whether such a signal depends on cognitive workload. Using a continuous virtual reality target-detection task, participants (N = 23) completed a within-subject workload manipulation (High vs. Low). At the team level, weighting votes by this pre-emptive neural signal, available before a response is committed, produced substantial accuracy gains on

arXiv:2609.02006v1 Announce Type: new Abstract: A compressed student has two shapes that need not agree: the weight it deploys at inference and the weight family its training can reach. We show that a state-of-the-art weight-inheritance distiller, Low-Rank Clone (LRC), deploys a full-width student MLP but ties training to a teacher-induced slice, leaving 62.5-81.4% of each deployed matrix's independent linear degrees of freedom unreachable-paid for at inference, never trainable. Our principle is one line: train what you deploy. From the identical LRC warm start, we make the training object the entire deployed matrix, with no change in deployed shape, deployed parameter count, or inference FLOPs, via two mergeable realizations (Dense-LRC and CORE-LRC) that both collapse to one deployed weight. This recovers stranded capacity: taking the stronger realization per teacher, +2.36/+2.71/+10.45 Avg9 over matched-budget plain-LRC baselines across three teachers (Llama3.2-3B, Llama3.1-8B,

A new report documents 592 silicosis cases among California engineered stone workers, including 31 deaths and 65 lung transplants since 2019.

The biggest healthy aging ‘blind spots’ — and doctors’ favorite biohacking tools for longevity  New York PostThis Simple Habit Could Help You Live Longer. But Does It Actually Work?  VICEU.S. News & World Report Survey: Health Experts Rank Core Lifestyle Habits Ahead of Biohacks for Longevity  MorningstarSpotlight: Redefining longevity: What does healthy aging actually mean?  Oakville NewsHere Are the Top Strategies for Healthy Aging, According to 63 Experts  Yahoo Health

The biggest healthy aging ‘blind spots’ — and doctors’ favorite biohacking tools for longevity  New York PostThis Simple Habit Could Help You Live Longer. But Does It Actually Work?  VICEU.S. News & World Report Survey: Health Experts Rank Core Lifestyle Habits Ahead of Biohacks for Longevity  MorningstarSpotlight: Redefining longevity: What does healthy aging actually mean?  Oakville NewsHere Are the Top Strategies for Healthy Aging, According to 63 Experts  Yahoo Health

Scientists study the underlying causes of a disease or test potential new drugs in animal models or cells obtained from human biopsy specimens before clinical trials. However, disease models in animals do not mimic actual human disease, and biopsy specimens are difficult to access for research purposes. Lab-grown tissues engineered from stem cells reprogrammed from human cells, called human induced pluripotent stem cells (hiPSCs), are rapidly emerging as a more reliable and accessible model in medical research.

Scientists developed tiny robots that use sound waves instead of motors to move, a design that could eventually benefit miniature medical devices.

The most common metabolic diseases, such as type 2 diabetes or fatty liver disease leading to metabolic dysfunction-associated steatohepatitis (MASH), a fibrotic condition of the liver, are consequences of the dysfunctional state of metabolism that results from being overweight and sedentary. These conditions are rare in thin, fit people, even in later life when the damage of aging makes metabolism more vulnerable and inflammatory. Excess visceral fat tissue is known to promote the accumulation of senescent cells. While cells become senescent throughout life, in youth they are promptly cleared by the immune system. That clearance falters with age, allowing senescent cells to linger and accumulate. The metabolic dysfunction produced as a consequence of being overweight accelerates this process. In today's open access paper researchers review […]

Scientists study the underlying causes of a disease or test new potential drugs in animal models or cells obtained from human biopsy specimens before clinical trials.

Insilico Medicine, a clinical-stage generative artificial intelligence (AI)-driven drug discovery company, today announced the release of a series of frontier specialist models for chemistry and biology, trained through its MMAI Gym for Science framework.

Laboratory experiments found that vanillin, a common vape flavoring, altered the development of human embryonic stem cells.

The recent flood of Artificial Intelligence (AI) models and content has undoubtedly brought along an unwelcome bug from the human-made content it scrapes from: bias. AI has been noted to form, copy, and amplify harmful stereotypes of already marginalized groups.

Giving old mice semaglutide—commonly sold as Wegovy or Ozempic—extends their lifespan remarkably, a new study finds

Stevie Bonifield / The Verge: MrBeast, whose YouTube channel recently hit 500M subs, will feature Gemini, Google Health, and the Fitbit Air in his videos as part of a multi-year Google deal  —  One of YouTube's biggest creators has a new deal to promote Gemini and Google Health.

Google Patents Wasnt Built for PharmaHeres What It Misses Every Time You Search a Drug Name Most people search Google the way they search for truth: type a name, hit enter, and assume the system will surface what matters. But if youre in pharmaregula... Source

AI Cant Be Named as an Inventor on Drug PatentsUntil Congress Says Otherwise What do drug developers, patent lawyers, and AI builders have in common right now? A growing mismatch between whats technologically possible and what the patent system legally... Source

Plus, news about Regeneron, Novocure, Jazz's Ziihera, BioMarin, Ascendis, AstraZeneca, Hutchmed and TScan Therapeutics. Another road bump for Tavneos: The Medicines and Healthcare products Regulatory Agency (MHRA) has followed the EMA’s lead ...

A study published in Nature has shown that the GLP-1 drug semaglutide extended the lifespan of older, healthy mice by tempering the detrimental effects of aging. Researchers at the University of California, Berkeley, directly compared the effects of the treatment with those of reduced food intake. They found that the drug mimicked the anti-aging benefits of calorie restriction, conferring even greater benefits in some areas.

The deaths of two children who received CRISPR therapies in China have rattled scientists who work on similar genetic medicines in the US, adding new fuel to a yearslong debate about the danger of the ...

The recent flood of artificial intelligence (AI) models and content has undoubtedly introduced an unwelcome flaw from the human-made content it scrapes: bias. AI has been shown to form, copy and amplify harmful stereotypes of already marginalized groups. This phenomenon has far-reaching effects as AI becomes a normal part of everyday life.

A federal judge has paused the Trump administration's abstinence-only teen pregnancy programme, citing unreliable AI-generated references. The ruling does not restore the £49.1m in cancelled grants.

Apple Watch знали, что я сделал. Obsidian знал, что я планировал. Но перед следующей тренировкой я всё равно смотрел на один красивый график и принимал решение почти вслепую. Рассказываю, как свёл план, фактическую нагрузку, самочувствие и медицинскую хронологию в обычные Markdown-файлы, а ввод отдал AI-агенту. Читать далее

After an organ transplant, a woman developed a new food allergy.

Cultivated meat aims to produce meat from animal cells in the laboratory without needing to raise a living animal for meat production. It has been proposed as a potentially more sustainable alternative to conventional meat production. Yet recreating something as complex as a steak—rather than minced meat—remains a major scientific challenge.

arXiv:2609.00430v1 Announce Type: new Abstract: Recent work uses large language models to generate systems code from specifications, treating the specification as the durable artifact and the implementation as disposable. Regenerating the implementation specializes it to each workload and device. However, that work lives in a forgiving setting: a component's externally visible effects, its writes and device commands, are recoverable, and the generator is honest, so trust is discharged by re-execution. We target the unforgiving setting: systems code whose effects are irreversible, produced by a generator that may be adversarial. There, re-execution cannot check an effect after the fact, and a proof fails silently when its assumptions do. We take the position that the only safe way to operate here is to deny the generated code the authority to act. The generated code only plans, while a fixed trusted mediator owns every effect and performs one only when the specification would have

arXiv:2609.00352v1 Announce Type: new Abstract: Large Language Models are now part of healthcare and mental health support systems, raising concerns regarding fairness toward vulnerable populations, including LGBTQIA+ individuals. However, limited empirical work has investigated how explicit LGBTQIA+ identity disclosure influences LLM-generated responses in mental health contexts. In this study, we extracted 50 real mental health questions from the Counsel Chat repository and constructed three prompt conditions for each question: no identity disclosure, explicit straight identity disclosure, and explicit LGBTQIA+ identity disclosure. We generated and analyzed 450 ChatGPT responses across these conditions using binary coding and comparative analysis. Our findings indicate that LGBTQIA+ identity disclosure did not substantially affect response completeness or supportive guidance. However, responses in the LGBTQIA+-explicit condition presented substantially more identity acknowledgment,

arXiv:2609.00319v1 Announce Type: new Abstract: Online health information seeking is shifting from keyword search, where users consider a ranked list of links, to conversational systems that compose a single answer and curate its citations. Source evaluation therefore passes from user to platform, yet what these systems surface is poorly characterized. We audited three free consumer products (ChatGPT, Perplexity, Google AI Overview) on twenty English mental health questions under two prompt conditions, with a subset of three also translated into six further languages of varying resource tiers. We recorded 15,942 citations across 1,140 responses and 1,713 unique domains, then classified every citation with a nine-category organizational typology applied by a deterministic classifier validated against human coding. Citations were heavily concentrated: the ten most-cited domains accounted for 43.6% of English citations, and government, commercial health, and academic sources were closely

Under the agreement, Andelyn will transition Genprex's adeno-associated virus (AAV)-based diabetes gene therapy candidate with process optimization, scale-up, and IND-enabling and clinical-stage manufacturing to support clinical trials. The post Andelyn to Support Manufacturing Scaleup for Genprex’s Diabetes Gene Therapy appeared first on GEN - Genetic Engineering and Biotechnology News.

Researchers today announced AdaptiveFlow, an AI-informed platform that can virtually screen billions of drug-like molecules with a 1,000-fold reduction in computational costs over existing methods. Developed and validated by scientists from St. Jude Children's Research Hospital, University of Pavia, Dana Farber Cancer Institute and Harvard Medical School, AdaptiveFlow allows prohibitively expensive ultra-large virtual drug screens to be conducted routinely. The open-source platform was published today in Nature Biotechnology.

Descriptive, Diagnostic, Predictive and Prescriptive—with two new domains: Critique and Creative. The result lays out where medical AI's greatest opportunities lie and how it could shift health care from treating disease to proactively reducing disease burden. At the same time, they emphasize that realizing this potential will require a careful balance between innovation, patient safety and responsible implementation.

In August, Google announced that Gemini users can now book appointments with Zocdoc. The news came not long after Solv said it’s partnering with Amazon ...

Senescent cells - cells that have stopped dividing but remain metabolically active - can accumulate with age and contribute to chronic inflammation and tissue dysfunction

Artificial intelligence isn't simply making healthcare more efficient—it's reshaping how healthcare organizations operate, compete, and serve patients. Author Tom Lawry joins Stewart Gandolf to explain why healthcare leaders must treat AI as a leadership challenge rather than a technology project.

A region of the brain called the striatum is critical for many cognitive and motor functions, including decision-making, movement control, habit formation and reward processing. It also plays a role in addiction and is significantly affected by Huntington's disease, schizophrenia and other disorders.

A new ethics paper from the American College of Physicians (ACP) says that while the use of artificial intelligence (AI) in medical practice can improve health care efficiency and quality, it must be grounded in clinical integrity and professionalism. The paper, published in Annals of Internal Medicine, outlines three guideposts for the ethical use of AI in medical practice: rationality, self-governance and competence.

Лечение показало эффективность у пациентов с наследственной гиперхолестеринемией и тяжелой гипертриглицеридемией – сложными состояниями, которые не поддавались контролю стандартными препаратами. После CRISPR-терапии пациенты достигли впечатляющих показателей по снижению симптомов, которые сохранялись не менее года.

RIVER FALLS, Wis. — Documentary filmmaker Peter Byck will discuss his documentary series, “Roots So Deep,” and present an abridged screening on Tuesday, Sept. 15, from 3 to 6 p.m. at the University of Wisconsin-River Falls in the Kinnickinnic River Theater in the University Center, 501 Wild Rose Ave. The event is free and open […] The post Documentary Filmmaker to Discuss Regenerative Grazing at UW-River Falls appeared first on Morning Ag Clips.

Every year, more Americans fly to Istanbul to deal with their hair loss and finally get a hair transplant in Turkey, and the reason they choose Turkey is not complicated. So people ask around, and a list of the top 10 hair transplant clinics in Turkey gets passed along. This is the 2026 version.

arXiv:2608.30656v1 Announce Type: new Abstract: Proactive tamper localization embeds an imperceptible signal into an image prior to distribution, enabling pixel-level manipulation detection. Existing methods assume a spliced (SP) setting, where synthesized regions are composited onto the original background, leaving embedded signals intact. However, real-world diffusion-based inpainting operates in a fully regenerated (FR) setting, where the entire image undergoes denoising, disrupting background signals and rendering existing frameworks ineffective. We propose APT, a semi-fragile latent-space perturbation that embeds a dense, vector-wise localization signal. By aligning each spatial feature vector toward a fixed anchor direction, APT localizes tampering via the alignment disparity between synthesized foreground and anchor-aligned background features after inpainting. The proposed hard negative mining loss and noisy perturbation branch further enforce uniform alignment. Experiments on

arXiv:2608.30369v1 Announce Type: new Abstract: We present OLIVE, a framework for adapting a foundation model to provide real-time assistance in temporally demanding, high-stakes, and dynamic tasks. We show that passive EEG, fused online with behavioral evidence, can meaningfully extend the number of targets users detect and engage beyond their unaided action bandwidth. OLIVE learns from both explicit behavioral signals (the targets the user shoots down in an XR first-person shooter game) and implicit physiological signals (fixation-locked EEG) to provide timely guidance, continuously adapting a frozen vision-language model's inference on which items are task-relevant by jointly estimating per-source reliability without manual labels or offline training. Through three user studies, including two live deployments of an assistive agent driven by OLIVE in XR, we show that OLIVE Pareto-dominates prior test-time adaptation frameworks, achieving the highest convergence rate at comparable

arXiv:2608.30130v1 Announce Type: new Abstract: Retrieval-augmented language models can fail to respect negative constraints when the retriever supplies evidence about concepts the user explicitly excluded. Beyond explicit negation, queries may ask for answers that include one concept while excluding another, or for entities that belong to a category but differ from a closely related instance. Because the excluded concept still appears in the query text, dense retrievers may assign high similarity to documents about that concept even when the user asks to avoid it. We introduce E-SENS, a training-free reranking method for negation-sensitive retrieval. E-SENS extracts a compact trap query for the excluded side and subtracts trap-query similarity from the original-query retrieval score. On ExcluIR, E-SENS shows a clear recall-violation trade-off across four embedding models and reduces trap retrieval at recall-preserving settings.

arXiv:2608.29965v1 Announce Type: new Abstract: Large language models can convert medical documents into structured data, but plausible output may still be unsupported by the source. Persisting such output in a longitudinal health record, a record that accumulates patient information over time, therefore creates an integrity risk: unverified data may influence later summaries, trends, or preventive-care computations. We introduce an evidence-gated trust-promotion model that keeps generated data provisional until a deterministic monitor verifies it against the source document. The monitor admits a candidate for a specified downstream use only when the source contains a unique supporting quotation, the relevant fields occur within the same laboratory row, and the required provenance is preserved. The generator cannot approve its own output, missing or ambiguous evidence causes refusal, and refused candidates remain available for human review rather than being silently discarded. We

arXiv:2608.29890v1 Announce Type: new Abstract: Biomedical Named Entity Recognition (NER) is fundamental to healthcare AI applications, including clinical decision support and medical information extraction. While corpora with Unified Medical Language System (UMLS) annotations, such as MedMentions, have driven progress in English biomedical NER, no comparable resource exists for Vietnamese. This paper presents En-ViMedNER, the first English-Vietnamese parallel biomedical NER corpus annotated with UMLS semantic types, which are language-neutral codes providing a shared cross-lingual label space and ensuring direct comparability with existing UMLS-based resources. The corpus contains 4,392 PubMed abstract pairs, 44,892 English-Vietnamese sentence pairs, and 202,949 aligned entity-mention pairs across 21 semantic types adapted from the MedMentions ST21pv dataset. To balance quality and scalability, we have constructed the corpus through automatic translation, expert post-editing,