AI News Archive: May 14, 2026 — Part 13
Sourced from 500+ daily AI sources, scored by relevance.
- Map2Soul Worldwide
The CBT, NLP Christian based Journaling and Learn app
- Smart Reply Pro
AI built for one specific job, to reply.
- The AI Business Automation Bible
Put your business on autopilot using ChatGPT & Claude.
- OptiProp
8 AI tools that write real estate content in seconds
- JobLine.ai
Shift Handoff Tracking and Manufacturing Dashboard
- Runverse
Your running coach that actually knows you.
- Framework
Think like a consultant with 100 AI-powered strategic models
- Rivvo AI
Your AI wingman for every text
- Bonito CLI
Deploy AI agents across any provider from one YAML file
- ClickBook — Offline AI Reader
Tap any word in books and understand it instantly offline.
- Treeova Technologies
No-code AI options trading automation for retail traders
- Fly Paper Web
Free SEO and AI search audit: see if ChatGPT & Google cite
- ClearAudit
AI website audit — know exactly why you're losing customers
- Magic Mind Reader
A clean, digital mind-reader that fits in your pocket.
- Azonova Sites
AI websites from docs, Maps, or Instagram
- Picksy
Picksy — безкоштовний AI-сервіс підбору фільмів серіалів
- Bonzi Studio - The 1st AI Editing studio
AI video generation, editing, and captions in one workflow
- aiHRly
Africa's #1 Native AI-Powered Applicant Tracking System
- Fonpilot
AI phone assistant – 24/7, zero setup, never miss a call
- Agent-Trust
Trust scores for autonomous agents on Base.
- StreamKit
Build and run Live video, speech-to-text, voice agent,
- Claude Code Skill Importer
Import Claude skills from .skill, .zip, SKILL.md, a folder
- TrySpeak.ai
Learn any language with AI conversation practice.
- The Promptory
The AI tools marketplace that finds your stack AND builds it
- Yoda A.I — Stocks · Crypto · Trading
Faster decisions with artificial intelligence.
- ListMagic
Generate perfect Etsy titles, tags & descriptions with AI
- ColorJibe
Ai Fashion Platform
- Alma by Olivares.AI
The workspace where memory, agents and creative studios share one canvas.
- Agentkit AI
No-code AI agent trained for your business.
- AI4Docs.AI
Clinical notes that write themselves in 100+ languages
- Retuner AI
Instantly tailor your resume to any job description.
- We and AI
AI companions that feel like real friends
- Stop Overthinking: Unlocking Efficient Listwise Reranking with Minimal Reasoning
Listwise reranking utilizing Large Language Models (LLMs) has achieved state-of-the-art retrieval effectiveness. Recently, reasoning-enhanced models have further pushed these boundaries by employing Chain-of-Thought (CoT) to perform deep comparative analysis of candidate documents. However, this per...
- Discrimination Is Generation: Unifying Ranking and Retrieval from a Tokenizer Perspective
Semantic IDs (SIDs) define the generation space of generative recommendation and directly determine its personalization ceiling. However, existing tokenizers are trained independently with retrieval objectives, leaving personalization signals fully decoupled from the SID construction process -- a fu...
- Asymmetric Generative Recommendation via Multi-Expert Projection and Multi-Faceted Hierarchical Quantization
Generative Recommendation (GenRec) models reformulate recommendation as a sequence generation task, representing items as discrete Semantic IDs used symmetrically as both inputs and prediction targets. We identify a critical dual-stage information bottleneck in this design: (1) the Input Bottleneck,...
- Efficient Generative Retrieval for E-commerce Search with Semantic Cluster IDs and Expert-Guided RL
Generative retrieval offers a promising alternative by unifying the fragmented multi-stage retrieval process into a single end-to-end model. However, its practical adoption in industrial e-commerce search remains challenging, given the massive and dynamic product catalogs, strict latency requirement...
- Towards Self-Evolving Agentic Literature Retrieval
As large language models reshape scientific research, literature retrieval faces a twofold challenge: ensuring source authenticity while maintaining a deep comprehension of academic search intents. While reliable, traditional keyword-centric search fails to capture complex research intents. Frontier...
- Pathway-Centric Integration of CRISPR Fitness with Molecular Features Draws Cancer State Maps
Cancer cells display heterogeneous pathway activity that shapes therapeutic vulnerability, but mapping it remains challenging. Transcriptomic scores do not directly measure functional activity, and CRISPR knockout data alone lack molecular interpretability. We introduce StateMap, a pathway-centric framework integrating gene expression and genome-wide CRISPR knockout fitness data from the Cancer Dependency Map. For a given pathway, StateMap selects features by co-dependency and mutual information, then projects cell lines into a low-dimensional space reflecting pathway activity and molecular state. Applied to the Hippo pathway, it resolved five functional states refining the YAP-on/YAP-off dichotomy. Notably, the 'Hippo-strong' state showed selective dependence on integrin V{beta}5; ITGAV depletion triggered Hippo-dependent cell aggregation and G1 arrest via enhanced cell-cell adhesion. Machine learning transfer to TCGA identified a matching subtype with poor prognosis, nominated NNMT a
- Explainable prediction and simulation of complex system dynamics through networks of manifolds
Complex systems such as brains and other interacting biological and physical processes are difficult to represent because they evolve across many variables, scales, and nonlinear interactions. To capture these multivariate, multiscale interactions we have developed Generative Manifold Networks (GMNs) a machine learning framework consisting of a network of linked dynamical systems. The network is discovered by an interaction function which can focus on causality, shared information, nonlinearity or other metric. Network nodes are low--dimensional data--driven state--space manifolds with generator functions accommodating multiscale dynamics. In contrast to many machine learning approaches GMNs have no latent or randomly initialized variables providing transparent explainability. GMNs generate short term dynamics of chaos on par with echo state networks while outperforming them in long term generation of chaos and neural dynamics, but with a markedly reduced number of dimensions and witho
- Classic machine learning on top of multiple position weight matrices improves genomic prediction of transcription factor binding sites
Motivation: DNA motifs recognised by transcription factors are typically represented as position weight matrices (PWMs), assuming independent contributions of individual nucleotides to protein binding specificity. Many alternative models accounting for correlations of positional contributions have been introduced in the past decades. However, performance gains have generally not out-weighed the advantages of simplicity, interpretability, and practical applicability of PWMs with the well-established codebase. Existing software tools and motif databases provide multiple non-identical PWMs for the same transcription factor or even for the same dataset. It remains a prac-tical question whether these PWMs can be effectively combined into a single improved model. Results: Here we describe ArChIPelago (https://github.com/autosome-ru/ArChIPelago), a compu-tational framework that combines multiple PWMs into a joint model using classic machine learning techniques, from linear regression to ensem
- OmniGene-4: A Unified Bio-Language MoE Model with Router-Level Interpretability
Mixture-of-Experts (MoE) architectures offer a rare opportunity to probe the internal organization of large language models, but this affordance has not been systematically exploited in biological foundation modeling. We introduce OmniGene-4, a unified bio-language foundation model built on Gemma-4-26B-A4B (30 layers, 128 experts per layer, top-8 routing) by injecting 28,028 biological tokens (DNA and protein BPE, Foldseek 3Di, DSSP secondary structure), continuing pretraining (CPT) on a 32.5 GB mixture of DNA, protein, natural-language and structural corpora, and supervised fine-tuning (SFT) on 199,576 instruction-format examples spanning eight task families. On a suite of standard benchmarks, the final model (v3) reaches 99.95% accuracy on BioPAWS standard protein homology (6,000 pairs), 59.50% on remote homology (2,000 pairs from protein_pair_remote), and 93.66% on BixBench knowledge questions. Relative to its un-fine-tuned vocabulary-extended Gemma-4-Instruct baseline (85% / 60% /
- Smartphone Placement Recognition during Walking: Performance Determinants and Real-World Generalizability
The opportunity to collect movement data from smartphones for prolonged periods has opened new perspectives in the field of clinical movement analysis. However, when monitoring people's mobility in free-living conditions, smartphone placement can influence the validity of the extracted digital mobility outcome. This study aimed to develop and validate an automatic smartphone placement recognition classifier and to investigate potential critical factors that can influence performance. The classifier was trained on data from 15 healthy participants using inertial signals collected from smartphones placed at six body placements during free-living walking and externally validated on over 3,000 individuals from external datasets, including blind participants and patients with cardiovascular or Parkinson's disease. A decision-tree ensemble model was developed using feature subsets of increasing dimensionality, with the optimal subset comprising 50 features. Classification accuracy increased
- Constrained Evolutionary Design of Matrixyl Analogs: Balancing Permeability and Functional Preservation Through Computational Optimization
Matrixyl (palmitoyl pentapeptide-4, KTTKS core) is a collagen-stimulating peptide used in topical anti-ageing products, but its in-use efficacy is limited by poor permeation through the stratum corneum. We describe a deterministic computational workflow that combines a tournament genetic algorithm and NSGA-II with exact RDKit molecular descriptors to search the fixed-length, edit-distance-2 neighbourhood of KTTKS (3,706 candidate sequences) for analogs with descriptors more favourable for passive transdermal diffusion. The search returns a 9-member Pareto frontier that quantifies the trade-off between predicted permeability and motif preservation. Five of the nine frontier members carry the same substitution, lysine to proline at position 4 (K4P). This single change lowers the topological polar surface area by 25.6%, removes the +1 charge contributed by lysine, and reduces the functional-preservation score from 1.00 (KTTKS) to 0.67. The frontier ranking is unchanged by +/-30% perturbat
- MagNet: Computational Methods for Constructing High-Confidence Protein-Protein Interaction Networks in Magnaporthe oryzae
Magnaporthe oryzae, the rice blast fungus, plays a role as a model organism for molecular plant-microbe interaction research. Studies on the pathogenic mechanism of this fungus revealed many genes involved in signaling pathways. As multi-omics data are being available, genomic-level researches have been conducted to uncover the underlying biological processes during the pathogenesis of M. oryzae. Identifying the genome-wide protein-protein interaction (PPI) network is one of the omics-level approaches, which helps to understand signaling and regulatory pathways. However, existing biological network resources of M. oryzae are not sufficient to decipher pathogenesis mechanisms due to the abundance of false positives/negatives. In this study, a reliable PPI network database of M. oryzae, MagNet, was constructed with three methods, including homology-based Interolog search, co-expression network construction, and domain-domain interaction (DDI)-based prediction. With three approaches altog
- STOMAPY: Artificial Intelligence for Risk Stratification of Outcomes Requiring Enterostomal Therapy After Hospital Discharge Following Colorectal Surgery
Introduction: Infectious and wound-healing complications after colorectal surgery often increase the complexity of local care and the need for specialized enterostomal therapy follow-up after hospital discharge. Despite the growing use of predictive models in digestive surgery, a translational gap remains between perioperative prediction and the practical organization of specialized care. Therefore, the aim of this study was to develop and temporally validate a machine-learning-based risk stratification model to estimate the probability of post-discharge outcomes associated with greater demand for enterostomal therapy after colorectal surgery. Methods: This was a retrospective observational study including 7,908 patients who underwent colorectal surgery between 2005 and 2014. The outcome was defined as the occurrence of superficial surgical site infection, delayed wound healing, or abdominal sinus formation. Routinely available preoperative and intraoperative variables were used as pre
- AnnotX: An Edge-powered Laparoscopic Video Annotation Platform
Accurate and objective evaluation of surgical skill and performance is critical for advancing training and improving patient outcomes. Current assessment methods increasingly rely on video analytics and depend on labor-intensive, frame-by-frame manual annotation by experts. In this work we developed a surgical video annotation platform (AnnotX) that used a Python backend running a pretrained promptable video segmentation foundation model, i.e., Segment Anything3 (SAM3) for per frame segmentation and temporal segment propagation. With a few interactions per class, the model generated a high-quality mask on a key frame and propagated it through the sequence. The platform automatically exported per-class binary masks and color overlays for every frame, together with deterministic metadata and a standardized study folder structure to support auditability and downstream analysis. On de-identified laparoscopic surgery videos, the system processed typical clips in minutes and reduced expert a
- US chip start-up Cerebras to raise $5.5 billion in IPO
US chip start-up Cerebras to raise $5.5 billion in IPO Gulf News
- Asymmetry between warmth and clinical substance in multilingual consumer health AI
The same patient question can yield different clinical quality across languages. Across 504 forum-derived patient queries in six languages and four chatbots, language-matched clinicians rated responses on five clinical dimensions (1,008 ratings; 5,040 dimension scores). Patient language outweighed chatbot identity across the four clinical-substance dimensions (composite language partial eta-squared 0.275 vs chatbot 0.035; robust to investigator-rating exclusion: eta-squared 0.260) but not for empathy (eta-squared 0.029): clinical substance was language-associated; warmth was relatively preserved. Catastrophic safety ratings ranged 4.3-fold by language (3.6% English, 15.5% Thai and Hebrew); 62% of catastrophic ratings exceeded the English baseline (descriptive disparity). Failures were systematic and silent: none of 24 stroke responses conveyed time-criticality framing, none of 24 CO-poisoning responses challenged the family's stress framing, and 120 sentinel responses contained no conf
- AI chipmaker Cerebras soars 90% in year’s biggest IPO so far
Artificial intelligence chipmaker Cerebras surged 90% in its trading debut, setting the bar for the biggest initial public stock offering of the year so far.
- Cerebras IPO: Pricing Is a Major Win for Chips; Stock Closes Up 68%
Cerebras IPO: Pricing Is a Major Win for Chips; Stock Closes Up 68% Barron's