AI News Archive: July 20, 2026 — Part 15
Sourced from 500+ daily AI sources, scored by relevance.
- FaceUnlock
Hands-free Mac unlock using your face - private, local, free
- Managed Hermes Agent | Hostinger
Hermes Agent with zero setup, maintenance, or infrastructure management.
- Kimi K3
Kimi K3
- Referent
Referent
- Borade AI
Borade AI
- Autonomous Discovery of Wireless Communications Algorithms
Large language model (LLM)-driven evolutionary search is an emerging algorithm-discovery paradigm that has already produced novel results in several scientific fields. Yet its application to wireless communications remains largely unexplored. To bridge this gap, we introduce The AI Telco Engineer (A...
- ETAS: An Effect-Typed Language for Agent Systems
ETAS is a programming language for agent systems that treats model-backed agents, tool calls, prompts, typed memory, human approvals, policies, and execution traces as semantic program elements rather than library conventions. It separates deterministic computation from agentic nondeterminism and ex...
- Lifelong Multi-Subsystem Pickup and Delivery with Buffer-Limited Handover Stations
Coordinating payload transfers between subsystems is a critical challenge in lifelong Multi-Agent Pickup and Delivery (MAPD). We study systems where agents are confined to separate regions and must exchange payloads through shared handover stations. These stations, equipped with single docks and fin...
- SR-Agent: An Experience-Driven Agentic Framework for Post-Ranking Strategies Refinement in E-Commerce Recommendation
User experience is a first-class objective in industrial e-commerce recommender systems (RS). Post-ranking strategies, which govern diversity, similarity, and exposure over a ranked list, are widely deployed in industrial RS for their simplicity and low serving cost. However, as the online recommend...
- I wanted it to feel more personal: Customization of social AI as AI individualism in practice
Despite the growing availability of customizable social artificial intelligence (AI), such as ChatGPT, Grok, and Character.ai, we know little about how users actively shape social AI to reflect their personal preferences. This study examines why and how users (N = 169) customize social AI through th...
- Persona-as-Configuration: Generative Stakeholder Reporting for Agricultural Floods
Cyber-physical systems built on deterministic edge inference, such as on-vehicle flood detection for agricultural fields, produce structured decision logs that must be interpreted differently by heterogeneous stakeholders. Pairing such systems with large language models (LLMs) to generate stakeholde...
- Informal Learning Emerges in Everyday Human-LLM Interaction
As LLMs become increasingly capable of completing tasks for users, a central concern is that everyday AI use may become primarily cognitive offloading, eroding the opportunities through which people develop their own capabilities. We analyse large-scale human--LLM conversations to ask whether inform...
- Human-in-the-Loop User Feedback Affects Perceived Accuracy and Trust, but Task Subjectivity Matters
While ML can produce complex models beyond those that a human could produce manually, incorporating human input can often improve performance beyond purely data-driven models. While this feedback could come from system designers or domain experts, in many cases, the end users who regularly use the s...
- Sidekick: Designing Communication for Effective Multitasking with Computer Use Agents
Computer Use Agents (CUAs) can autonomously execute complex, multi-step tasks within GUIs, enhancing efficiency through parallel multitasking. However, our formative studies with CUA experts and GenAI users indicated that current feedback is primarily text-based, requiring sustained attention to mon...
- HyCoRec: Hypergraph-Enhanced Multi-Preference Learning for Alleviating Matthew Effect in Conversational Recommendation
The Matthew effect is a notorious issue in Recommender Systems (RSs), \emph{i.e.}, the rich get richer and the poor get poorer, wherein popular items are overexposed while less popular ones are regularly ignored. Most methods examine Matthew effect in static or nearly-static recommendation scenarios...
- RRAM-DP: Device-Calibrated Differential Privacy for In-Memory Edge Learning
Edge Artificial Intelligence of Things (AIoT) systems often collect sensitive data in situ, raising serious privacy concerns. Resistive-switching random-access memory (RRAM) is an attractive substrate for efficient AIoT thanks to its multi-bit storage and compute-in-memory (CiM) capabilities, while ...
- CutBackdoor: A Circuit Cut Triggered Backdoor Attack on Variational Quantum Algorithms
Variational Quantum Algorithms (VQAs) are a leading paradigm for near-term quantum computing, combining parameterized quantum circuits with classical optimization across quantum chemistry, combinatorial optimization, and quantum machine learning. Since real-world VQA deployments routinely require ci...
- GARAGE: Characterizing the Automation Boundary in LLM-based Attack Graph Generation
While modern vehicle security depends on effective Cyber Threat Intelligence (CTI) synthesis, current automated tools struggle with unstructured data and automotive-specific architectural nuances. To bridge this gap, we introduce GARAGE, a RAG-powered framework that converts fragmented CTI into an a...
- Residual Observability and Attack Detectability in Encrypted OPC UA Traffic
OPC Unified Architecture (OPC UA) encryption conceals application-layer semantics and restricts intrusion detection to residual communication structure. Although machine learning-based intrusion detection systems (IDSs) can detect attacks in encrypted OPC UA traffic, the relationship between residua...
- Insecure Coding Preferences in Long-Term Memory: Security Risks for LLM-based Code Generation
LLM-based systems increasingly incorporate long-term memory to improve cross-session continuity. However, once insecure coding preferences are stored, they may silently influence security-critical decisions in subsequent generations. In this study, we conduct the first systematic empirical study on ...
- Protecting Floating-Point Computation for DNN Binaries with MBA Obfuscation
Deep neural networks (DNNs) have become a foundational component of modern computing systems with a wide range of applications, such as computer vision, edge intelligence, etc. For the sake of low latency and data privacy, DNN models are increasingly compiled into executables and deployed on local d...
- (A)iSpy: Parasitic Trojans for Machine Learning Infrastructure
Modern machine learning (ML) pipelines depend heavily on third party libraries for graph compilation and hardware acceleration. While current practices audit data and model artifacts or rely on file integrity checks, the execution environment remains implicitly trusted. This blind spot enables activ...
- ShadowPickle: Evading Machine Learning Model Scanners via Stealthy Pickle Deserialization Attacks
Model hosting hubs (e.g., Hugging Face) are vulnerable to supply chain attacks that enable remote code execution on trusted user environments. Attackers often distribute malicious Pre-trained ML models (PTMs) via model hubs. In this paper, we present novel attacks against PTMs and model hubs called ...
- MeshScope-Region: Distribution, Road-Network Accessibility, and Nine-Year Evolution of ICU and HCU Capacity Across Japan's 330 Secondary Medical Areas
Background: In Japan, health planning is organized around secondary medical areas (SMAs; niji-iryo-ken; 330 areas in the 2025 classification), yet nationwide analyses of intensive care unit (ICU) capacity have been conducted mainly at the prefecture level, and a recent SMA-level study addressed only the presence or absence of ICUs. The full supply structure of intensive and intermediate critical care - ICU and high care unit (HCU) beds - has not been characterized at the SMA level with respect to its composition, road-network accessibility, and evolution over time. Methods: We developed MeshScope-Region, an analytical platform built on the Hospital Bed Function Reports (byosho-kino-hokoku) for fiscal years 2016-2024, in which ICU and HCU beds were identified from notified reimbursement categories and aggregated to SMAs. Three analytical layers were integrated: (1) cross-sectional distribution of ICU/HCU beds; (2) nationwide road-network accessibility computed with the Open Source Routing Machine (OSRM) from 176,962 populated 1-km census grid cells to all facilities reporting ICU or HCU beds; and (3) a nine-year longitudinal analysis of supply-structure types, classified by k-means (k = 6) in an 8-dimensional PCA space anchored to fiscal year 2024, with earlier years projected into the same space. Results: In fiscal year 2024, 20,631 ICU/HCU beds were reported nationally (7,114 ICU-type; 13,517 HCU-type) at 1,044 facilities. Zone-level totals among SMAs with any beds ranged 229-fold (3-688 beds); the 90th/10th percentile ratio of per-capita density was 3.6. In total, 90.1% of the population resided within 30 minutes' drive of a facility with ICU beds and 97.8% within 60 minutes; only 0.8% resided beyond 90 minutes. Although 140 of the 330 SMAs had no ICU facility within their own boundaries, 84.7% of their residents could reach an ICU facility in an adjacent area within 60 minutes' drive. Longitudinally, supply structures were highly persistent: 63.0% of SMAs (208/330) retained the same structural type across all nine years, adjacent-year rank correlations of a supply-vulnerability index were 0.887-0.924 (2016 vs. 2024: rho = 0.711), and the number of SMAs with zero ICU beds remained frozen at 133-141. The Gini coefficient of bed distribution declined from 0.384 to 0.262 - although computed on ICU-type beds alone it remained 0.365 in fiscal year 2024 - and capacity growth (total +27.9%) was driven predominantly by HCU beds (+41.6%) while ICU beds grew only +8.0%. Conclusions: Japan's critical care supply structure is regionally rigid, with a stable set of approximately 140 SMAs lacking ICU beds for nearly a decade, yet road-network accessibility substantially mitigates the consequences of zone-level absence. Recent capacity growth - and much of the apparent equalization - has occurred predominantly in intermediate care. MeshScope-Region provides a standing, reproducible evidence base at the geographic unit of Japan's medical planning cycles.
- Detection, Attribution, Narration: An End-to-End Pipeline for Explainable Money Mule Identification
Money mule accounts are critical facilitators of financial fraud, yet detecting them at scale remains challenging due to the heterogeneous nature of transactional and behavioural data. We present an end-to-end pipeline for customer-level mule detection comprising three stages: (1) a LightGBM classif...
- Testing Retrieval-Augmented Generation Systems with Chunk Coverage
Retrieval-Augmented Generation (RAG)-based systems\footnote{For brevity, RAG-based systems are referred to as RAG systems throughout this paper.} are increasingly deployed in high-stakes settings where correct behaviour depends not only on the language model but also on the retrieval component that ...
- DepRepair: LLM-Based Source-Code Repair for Dependency Breaking Changes
Modern software projects depend on numerous third-party libraries, whose updates often introduce breaking changes. Adapting consumer code to such changes remains labor-intensive and error-prone. Existing work either characterizes dependency breaking changes without producing a verified consumer-side...
- KernelDiag: Agent-Based Root Cause Diagnosis for Kernel Crashes
The Linux kernel is one of the most complex software systems, where automated fuzzing continuously exposes thousands of crashes, yet root-cause diagnosis remains a manual and time-consuming bottleneck. Existing LLM-based root cause analysis (RCA) techniques, effective for distributed systems, do not...
- Verify, Repair, Repeat, or Stop? Robust Stopping for Noisy Verify-Repair Loops in LLM Agents
Verify-repair loops are a standard means for large language model (LLM) agents to correct faulty plans in code generation, mathematical reasoning, and tool use. When both the verifier and the repairer are noisy, repair can damage already-correct plans, and reported acceptance keeps rising while true...
- FailureAtlas: A Taxonomy of Failure Modes in Multi-Provider LLM Serving Infrastructure
Multi-provider LLM gateways reverse proxies that route, load-balance, and rate-limit requests across foundation-model APIs have become critical production infrastructure. Yet the failure modes specific to this architectural layer remain undocumented, scattered across issue trackers and post-mortems ...
- Test Coverage Analysis of Agentic Pull Requests
AI coding agents increasingly submit complete pull requests (PRs) with minimal human intervention, shifting software development from AI-assisted to autonomous workflows. As these agents become more prevalent, ensuring the code they generate is adequately tested, by existing tests or by tests the ag...
- How Agent Skills Fail under Long Contexts: A White-Box Study in Code Auditing
Agent Skills package procedural instructions and checks for use by general-purpose agents, but loading a skill does not guarantee that every requirement remains active throughout a long tool-using trajectory. We study this problem in a production-derived, white-box code-audit workflow. Holding the t...
- (Over)Reliance on Test Agents in AI-Assisted Software Testing
AI-based test agents promise to accelerate software testing by shortening feedback loops in continuous development and improving scalability and maintainability. To realize these benefits, engineers must still be able to assess if agent outputs are useful, valid, and reliable, rather than treating t...
- CommitLLM: A Fine-Tuned Pipeline for Git Commit Message Generation
Developers frequently write uninformative git commit messages such as "fix" or "update stuff", degrading the value of version-control history for code review, debugging, and onboarding. We present CommitLLM, a three-stage pipeline that generates concise, Conventional Commits-compliant messages from ...
- The tttAI System for the TSA-ASR Task of the SmartGlasses Challenge 2026
This paper presents the tttAI system submitted to the TSA-ASR task of the SmartGlasses Challenge 2026, evaluated on both two-person dialogues (Track 1) and multi-party meetings (Track 2). The task requires time-stamped speaker-attributed speech recognition from smart-glasses recordings. This is part...
- X-Translator: A Real-Time Multilingual Speaker-Aware Speech-to-Speech Translation System
Real-time speech-to-speech translation (S2ST) systems must balance translation quality, latency, speech naturalness, and speaker consistency. Publicly documented S2ST systems have advanced direct, multilingual, streaming, and expressive modeling, while proprietary products and APIs increasingly expo...
- Evidence-in-the-Loop: Trace-Driven Optimization for Customer-Service LLM Agents
Production customer-service bots must improve answer quality across iterative releases, yet large language models must not bypass evidence boundaries, policy rules, or human-handoff safeguards. We present an \textbf{Evidence-Grounded Customer-Service Agent Workflow} deployed in a real-world customer...
- ANNLib: A Development Framework for Efficient Approximate Nearest Neighbor Search
Approximate Nearest Neighbor Search (ANNS) plays a pivotal role in modern deep learning pipelines. Recently, many ANNS systems have been proposed to either provide broad functionality or reach high performance. However, it is yet difficult to achieve both with minimal programming efforts. We propose...
- The Matryoshka Hypencoder
The Hypencoder is a recently-proposed retrieval approach that encodes queries as shallow neural networks ("Q-Nets") that estimate relevance over pre-computed document embeddings. Inspired by Matryoshka Representation Learning, we show that the Hypencoder can be extended to support multiple sizes of ...
- jina-reranker-v3.5: An Efficient Listwise Reranker with Hybrid Attention and Self-Distillation
Listwise rerankers are the discriminative core of agentic retrieval pipelines, yet production deployment demands efficiency, domain robustness, and fluency on semi-structured data at the same time. We present jina-reranker-v3.5, a 0.6B-parameter listwise reranker that meets these demands together wi...
- MagicSelector: Joint Optimization for Agent Tool Selection via Counterfactual Decomposition and Progressive Reranking
We present MagicSelector, a joint optimization framework integrating Counterfactual task decomposition, Progressive reranking, and Dynamic Top-K, designed to address the fundamental challenges of tool retrieval in agents. MagicSelector is a specialized framework capable of translating ambiguous user...
- RAMP: Robust Ad Recommendation Under Limited Personalized-Feature Availability via Masking and Alignment Pathways
Click-through rate (CTR) and conversion rate (CVR) prediction are fundamental tasks in online advertising, aiming to estimate the likelihood of user interactions based on various features. While personalized attributes such as age and gender can significantly enhance predictive accuracy, their use i...
- GCM: metric-guided clustering by genetic algorithm for correlation-defined modules
Gene co-expression analyses identify "good" modules by a correlation criterion. However, standard pipelines detect modules with greedy algorithms that optimize other quantities and only measure correlation afterwards. We present a method called Genetic Clustering by Metric (GCM), an open-source Python tool that closes this gap by treating module detection as maximum-likelihood inference and solving it globally. In GCM, the correlation objective is, up to a constant and the sample-size factor, the profile log-likelihood of an explicit generative model: a block-diagonal one-factor Gaussian in which each module is a single regulator with equal-magnitude loadings. This bases model selection on a principled footing through a genuine BIC/AIC in correlation space. GCM maximizes this likelihood with a memetic genetic algorithm: a population-based search hybridized with a greedy local refinement that reassigns genes after the fact, a move the agglomerative clustering at the core of co-expression pipelines cannot make. Across a replicated noise sweep, GCM reproducibly surpasses hierarchical correlation clustering and k-means with the lowest variance, and an ablation shows the local-search step is responsible; the advantage persists when the number of modules is unknown and when unstructured genes must be ignored. GCM faithfully optimizes geometric indices on the Iris benchmark dataset. For a breast-cancer RNA-seq it recovers coherent modules that predict tumor-versus-normal status.GCM depends only on NumPy and SciPy and exposes one swappable-metric interface with single- and multi-objective modes.
- scRepresenter: a workflow for computing, integrating and benchmarking cellular representations in single-cell transcriptomics
Motivation: Single-cell RNA sequencing (scRNA-seq) has become an attractive tool for studying complex diseases, in which transient cell states affecting diverse cell populations characterise disease development and progression. However, due to data sparsity and disease heterogeneity analysis is often challenging. With recent advances in machine learning, two widely used approaches have emerged for learning cellular representations: large-scale foundation models and biological knowledge-guided methods. Despite their complementary strengths, there is currently no unified workflow for systematically comparing and integrating these approaches. Results: Here, we present scRepresenter, an open-source workflow for computing, integrating, and validating cellular embeddings derived from foundation models and biological knowledge-guided methods in the context of complex diseases. It consists of two components: a command-line workflow that computes cellular embeddings and performs downstream analyses, and an interactive Shiny application for visualizing and comparing the computed embeddings. scRepresenter supports four categories of cellular representations: (1) expression-based, (2) knowledge-guided, (3) foundation model-derived, and (4) hybrid embeddings that combine foundation model-derived representations with knowledge-guided representations. This approach takes a cell-by-gene count matrix as input and outputs an integrated object containing the computed embeddings. Then, this object can be uploaded into our interactive Shiny application to compare different embeddings. Availability: The workflow is available at https://github.com/GuilhermePocas/scRepresenter Contact: AL291@cam.ac.uk; MA2129@cam.ac.uk Keywords: single-cell RNA-sequencing, machine learning, representation learning, foundation models, cellular embeddings, complex diseases, amyotrophic lateral sclerosis, human organoid
- Toward routine health phenotyping: High-throughput prediction of metabolic, immune, and inflammatory biomarkers from milk mid-infrared spectroscopy in early-lactation dairy cows
This study evaluated the potential of milk mid-infrared (MIR) spectroscopy, combined with routinely available on-farm variables, for predicting serum metabolic, immune, and inflammatory biomarkers in early-lactation cows. Data included 5,936 blood samples from 4,442 cows across 23 Australian dairy herds, with paired milk MIR spectra and serum measurements for up to 14 biomarkers. Prediction models were developed using partial least squares regression and evaluated using nested 10-fold random cross-validation and leave-one-herd-out validation. The results show that while basic herd-test data, including milk fat, protein, and lactose concentration, as well as on-farm variables, including DIM, calving age, breed, and herd could predict serum biomarkers, combining MIR spectra with these on-farm variables produced the best overall performance. In random cross-validation, blood urea nitrogen (BUN) was predicted most accurately (R2 = 0.78), while {beta}-hydroxybutyrate (BHB) and nonesterified fatty acids (NEFA) showed moderate accuracy (R2 = 0.56 and 0.44, respectively). BUN also showed the strongest external validation performance, with leave-one-herd-out R2 = 0.58 and comparable accuracy for predicting records collected after 70 days in milk (R2 = 0.65). BHB and NEFA had moderate leave-one-herd-out accuracy but did not transfer beyond early lactation. Most other biomarkers showed low or inconsistent external validation performance. Overall, MIR spectroscopy combined with on-farm variables shows promise for routine prediction of BUN, BHB and NEFA, which can be used for monitoring and genetic evaluation of, for example, ketosis and energy deficit. Initial random cross-validation results for glucose, bilirubin and cholesterol were promising, but more data is needed to improve the prediction accuracy and robustness of the predictions.
- Travel health needs in people visiting friends and relatives: a retrospective analysis of the UK National Travel Health Advice Line, 2019-2025
Background: Travellers visiting friends and relatives (VFRs) experience a disproportionate burden of travel-associated infectious diseases, yet little is known about the complexity of pre-travel consultations required to support their care. We compared enquiries relating to VFR travellers and tourists received by the UK National Travel Health Network and Centre (NaTHNaC) specialist Advice Line to identify differences in traveller characteristics, destinations and clinical complexity. Methods: We conducted a retrospective observational study of enquiries to the NaTHNaC Advice Line between 1 January 2019 and 31 December 2025. Enquiries relating to VFR travellers and tourists were compared using descriptive statistics and appropriate statistical tests. Traveller demographics, travel characteristics, destinations and enquiry management were analysed. Results: Of 16,367 enquiries relating to specific travellers, 3,090 (18.9%) concerned VFR travellers and 7,237 (44.2%) concerned tourists. Compared with tourists, VFR travellers were younger (median age 24 vs 52 years, P<0.001), more likely to undertake long-stay (8.4% vs 1.8%, P<0.001) and last-minute travel (5.0% vs 1.1%, P<0.001), and more frequently travelled to the WHO African Region (56.6% vs 29.2%, P<0.001) and Eastern Mediterranean Region (12.7% vs 2.8%, P<0.001). Pregnancy was substantially more common among VFR travellers (11.6% vs 4.3%, P<0.001). Enquiries concerning VFR travellers were more likely to require a call-back (16.1% vs 13.8%, P=0.012) and escalation to a specialist doctor (13.1% vs 10.5%, P<0.001), indicating greater consultation complexity. General practice generated a higher proportion of VFR-related enquiries than tourist enquiries (69.5% vs 63.9%, P<0.001). Conclusions: VFR travellers generate disproportionately complex pre-travel consultations characterised by higher rates of specialist escalation, distinct travel patterns and travel to destinations associated with the greatest burden of imported infectious diseases. These findings highlight the importance of specialist travel medicine support for healthcare professionals managing VFR travellers and reinforce the need for equitable access to timely, high-quality pre-travel healthcare for this high-risk population.
- NUMonomer enables accurate and scalable nucleic acid structure prediction from primary sequence alone
Accurate and efficient prediction of three-dimensional nucleic acid structures can accelerate functional characterization and enable downstream applications. Recent deep-learning methods have substantially improved nucleic acid structure prediction by incorporating auxiliary inputs such as multiple sequence alignments, secondary-structure annotations, and representations from pretrained language models. However, prediction accuracy remains limited, and generating these auxiliary inputs can be computationally expensive. Here we show that learning the hierarchical organization of experimentally determined structures across multiple scales, from recurring local conformations to global fold topologies, together with exploiting representations shared between RNA and single-stranded DNA, improves model generalization. Guided by these findings, we developed NUMonomer, an end-to-end deep-learning framework trained with input sequences spanning thousands of nucleotides on a joint RNA and single-stranded DNA dataset to predict nucleic acid structures directly from sequence. Despite requiring no auxiliary inputs, NUMonomer matches or outperforms leading prediction methods on benchmarks comprising CASP16 RNA targets and non-redundant sets of experimentally determined RNA and single-stranded DNA structures, with particularly pronounced improvements for longer RNAs. Its efficient and scalable architecture also reduces inference costs by approximately two orders of magnitude relative to the evaluated methods, enabling large-scale structure prediction. Together, these findings provide insight into generalization in biomolecular structure learning and establish NUMonomer as a practical framework for nucleic acid structure prediction.
- MedZone Embedder: a framework for representation learning of Japanese secondary medical care areas from a national ICU registry, characterizing intensive care provision structure and regional vulnerability
Background: In Japan, acute inpatient care is divided into approximately 335 secondary medical care areas, which serve as the basic units for planning healthcare delivery systems under the 8th National Health Care Plan. While comparisons between regions and facilities typically rely on a single risk-adjusted metric, this approach confuses differences in patient demographics with differences in the actual infrastructure of intensive care units (ICUs). This paper presents a framework - MedZone Embedder - for deriving data-driven indicators of regional structural vulnerability by mapping secondary medical care areas onto a learned similarity space, together with its working implementation. The paper sets out the concept, the method, a proof of concept, and an explicit staged validation program, rather than national empirical results. Methods: Each area is represented by a feature vector consisting of aggregated values of intensive care provision indicators derived directly from the Japan Intensive Care Patient Database (JIPAD) - specifically, risk-adjusted mortality rates (standardized mortality ratios and an in-hospital composite indicator), technical efficiency, length of stay, readmission rates, case severity, and case composition - with the within-area variance of these indicators also taken into account. No hierarchical processing by facility type is performed. A contrastive autoencoder (multilayer perceptron encoder 32 -> 16 -> 8, symmetric decoder) is trained by self-supervised learning, using an objective function that combines reconstruction and normalized temperature cross-entropy (NT-Xent) on noise-augmented views. The resulting 8-dimensional embedding supports area searches based on cosine similarity and anomaly scoring in the embedding space (using isolation forest, Mahalanobis distance, or k-nearest-neighbor density), which is normalized to a vulnerability score ranging from 0 to 1. If deep learning libraries are unavailable, or if the number of areas is small, an alternative method using deterministic principal component analysis is employed. Results: This method was implemented and deployed within an operational ICU decision support system on a managed cloud platform. The proof of concept (PoC) is structured around five secondary medical care areas within Kyoto Prefecture and runs entirely on synthetic facility-level aggregate data constructed to follow the JIPAD indicator schema; no registry data were accessed. It generated: an aggregate provision profile for each area; an area embedding space equipped with a similar-area search function; and a vulnerability ranking that identifies areas with low patient numbers and low diversity that exhibit overall poor outcomes. At this scale, the contrastive autoencoder falls back to principal component projection. The deep learning pathway has been implemented and unit testing has been completed; training and evaluation on actual registry data are pending data-use approval and the expansion of data integration. Validation is staged: Stage 2 will train the contrastive pathway over JIPAD-covered areas to assess construct validity against public structural indicators (ICU/HCU beds, population, accessibility), and Stage 3 will extend coverage to all areas via National Database (NDB) linkage. Conclusion: MedZone Embedder reframes regional comparison from single-indicator ranking to structural representation: which areas are alike, and which are structural outliers. The contribution of this paper is the framework - the proposal that the intensive care provision structure of Japanese secondary medical care areas can be learned from a national outcomes registry and read through the lens of what we call institutional debt - together with a deployed implementation and a pre-specified validation program. To our knowledge, this is a candidate first application of contrastive representation learning to Japanese secondary medical care areas.
- Diagnostic Accuracy of MRI Radiomics for Predicting KRAS Mutation in Rectal Cancer: A Systematic Review and Meta-analysis
Background: KRAS mutation status is an important biomarker in rectal cancer, with implications for prognosis and treatment response. MRI-based radiomics has emerged as a non-invasive approach for predicting tumor genotypes. However, the diagnostic performance of MRI radiomics for predicting KRAS mutation status remains unclear. This study aimed to evaluate the diagnostic accuracy of MRI radiomics for predicting KRAS mutations in rectal cancer. Methods: A systematic search of PubMed, Cochrane Library, Scopus, and Web of Science was performed through July 2025. Diagnostic test accuracy studies evaluating MRI-based radiomics or artificial intelligence models for predicting KRAS mutation status in adult patients with rectal cancer were included, using molecular testing as the reference standard. Risk of bias was assessed using the QUADAS-2 tool. Pooled sensitivity and specificity were estimated using a bivariate random-effects model. Results: Seven studies involving 1,224 patients were included. The pooled sensitivity was 0.736 (95% CI: 0.697-0.772) and the pooled specificity was 0.645 (95% CI: 0.586-0.701). The false positive rate was 0.355 (95% CI: 0.299-0.414). The area under the hierarchical summary receiver operating characteristic curve was 0.754, with a normalized partial AUC of 0.666. Between-study heterogeneity ranged from low to moderate depending on the estimation method (I2 = 8.4%-53.3%). Conclusion: MRI radiomics demonstrates moderate diagnostic accuracy for predicting KRAS mutation status in rectal cancer and may serve as a promising non-invasive biomarker for preoperative molecular stratification. Further large-scale studies with external validation are required to confirm its clinical utility.
- Diffusion MRI Profiles Map onto Distinct Inflammatory States After Adolescent Concussion: A CARE4Kids Study
Importance: Neuroinflammation is a key component of the response to injury after concussion, but direct links between diffusion MRI metrics and specific plasma inflammatory pathways in human concussion have not been established. Objective: To examine associations between diffusion MRI metrics and pathway-level inflammatory proteomic signatures in adolescents during the subacute period after concussion. Design, Setting, and Participants: Cross-sectional analysis of data from the CARE4Kids Consortium, a six-site prospective study. Participants were English-speaking adolescents ages 11-17.99 with concussion and symptoms at 7-35 days post-injury. Data were collected between 2022-2024. Of 370 enrolled participants, 122 had both diffusion MRI and plasma proteomics available for analysis. Exposure: Advanced diffusion MRI metrics were converted to z-scores and participants were grouped by the spatial extent of outlier values (potholes and peaks) across 15 white matter regions of interest. Nine non-redundant groupings were selected for primary analysis. Main Outcomes and Measures: Pathway-level inflammatory profiles derived from gene set enrichment analysis (GSEA) of ~5,400 plasma proteins measured by Olink proximity extension assay, targeting nine hallmark inflammatory pathways spanning initiation through resolution. Persistent symptoms were assessed 64-115 days post-injury. Results: Diffusion metrics reflecting tissue disorganization were associated with upregulation of the coagulation pathway, consistent with hemostatic-inflammatory signaling. Metrics reflecting reduced tissue complexity and neurite density were associated with upregulation of interferon- and interferon-{gamma} response pathways, consistent with microstructural remodeling driven by cellular immune activation. Elevated free water content was associated with downregulation of most inflammatory pathways and trend-level transforming growth factor - {beta} upregulation, reflecting inflammatory resolution. Time since injury did not differ between groups based on free water (Kolmogorov-Smirnov p = 0.97), suggesting these differences reflect individual variability in recovery pace. Exploratory analyses showed a trend toward lower odds of persistent symptoms in the group with elevated free water content (odds ratio = 0.51, p = 0.18). Conclusions and Relevance: Multiple diffusion MRI metrics are differentially sensitive to distinct neuroinflammatory states in the subacute period after adolescent concussion. These findings suggest that diffusion imaging could serve as a non-invasive tool for inflammatory phenotyping, with potential implications for identifying patients who may benefit from targeted immunomodulatory intervention.