paper_id string | title string | published_date string | arxiv_categories list | legal_license string | license_url string | commercial_use_allowed bool | commercial_ip_safety_score int64 | commercial_ip_verdict string | commercial_ip_justification string | reproduction_difficulty_level int64 | reproduction_difficulty_tier string | academic_citations_count int64 | field_weighted_citation_impact_fwci float64 | primary_author string | total_authors_count int64 | authors list | affiliations list | arxiv_url string | code_audit_status string | primary_repo_url string | repo_urls list | github_stars int64 | github_forks int64 | github_last_commit string | github_license string | importance_score float64 | abstract string | title_vector_384d list | abstract_vector_384d list | target_industry string | execution_mode string | model_parameter_scale string | min_inference_vram_fp16_gb float64 | min_inference_vram_quantized_gb float64 | recommended_gpu_class string | serving_frameworks_compatible list | tested_benchmarks list | extracted_benchmark_scores list | limitations_and_tradeoffs list | reproduction_recipe string | core_problem_addressed string | key_technical_innovation string | key_quantitative_result string | trend_velocity_tier string | papers_last_30_days int64 | audited_at string |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2607.13940v1 | A Self-Evolving Agent for Longitudinal Personal Health Management | 2026-07-15T15:22:11Z | [
"cs.AI"
] | ArXiv Standard | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | true | 80 | Enterprise Safe (Commercial Training & Deployment Allowed) | ArXiv Standard Distribution; Permissive Open-Source Software (MIT) | 1 | Level 1: Plug-and-Play (Verified Package & Checkpoint Available) | 0 | 1 | Haoran Li | 13 | [
"Haoran Li",
"Jiebi Deng",
"Tong Jin",
"Jinghong Han",
"Yuxin Wang",
"Zexin Wang",
"Qingyi Si",
"Weikang Gong",
"Xiahai Zhuang",
"Jia You",
"Wei Cheng",
"Jianfeng Feng",
"Hongcheng Guo"
] | [
"Clinical / Biomedical AI Research Institute"
] | http://arxiv.org/abs/2607.13940v1 | VERIFIED_LIVE | https://github.com/HC-Guo/HealthClaw | [
"https://github.com/HC-Guo/HealthClaw"
] | 147 | 6 | 2026-07-27 | MIT | 82.05 | Personal health management unfolds over repeated encounters, yet most health AI systems treat each request in isolation. We developed HealthClaw, an open-source agent architecture that updates support as a person's routines, preferences, measurements and risks change. It separates shared safety rules and medical knowle... | [
-0.01430600043386221,
0.002428000094369054,
-0.025133000686764717,
0.058024998754262924,
0.00496899988502264,
-0.007513000164180994,
0.02024500072002411,
0.0644569993019104,
0.026510000228881836,
-0.07679499685764313,
0.012264999561011791,
-0.016119999811053276,
-0.006146000232547522,
-0.0... | [
-0.04007500037550926,
0.022400999441742897,
-0.07395000010728836,
0.03409399837255478,
0.005872999783605337,
0.009198999963700771,
-0.05659300088882446,
0.07632599771022797,
-0.04327699914574623,
-0.06298600137233734,
-0.03528499975800514,
0.0008040000102482736,
-0.004476000089198351,
-0.0... | Medical AI, Clinical Foundation Models & Healthcare Intelligence | Clinical Decision Support & Algorithmic Pipeline | Model-Agnostic / Clinical Architecture | 0 | 0 | CPU Server / Model-Agnostic Clinical Host | [
"MONAI Medical Framework",
"Custom PyTorch Clinical Pipeline",
"FHIR / HL7 Adapter"
] | [
"Medical Imaging & Multi-Center Diagnostic Radiography"
] | [
"Quantitative Clinical Diagnostic & Statistical Evaluation"
] | [
"Data Privacy & HIPAA / GDPR Boundaries"
] | git clone https://github.com/HC-Guo/HealthClaw && cd HealthClaw && (pip install -e . || pip install -r requirements.txt) | Personal health management unfolds over repeated encounters, yet most health AI systems treat each request in isolation. | We developed HealthClaw, an open-source agent architecture that updates support as a person's routines, preferences, measurements and risks change. | Across 900 longitudinal support probes, answer accuracy increased from 0.2% with current-query prompting to 45.7% with HealthClaw, while prompt-side context exposure was 71.7% lower than with full-history prompting. | Explosive (>50/mo) | 192 | 2026-08-18T15:17:48.738653 |
2607.27235v1 | RadHarmony: Radiological Data Handling in the Era of Agentic AI | 2026-07-24T15:25:26Z | [
"cs.AI",
"cs.CV"
] | ArXiv Standard | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | true | 80 | Enterprise Safe (Commercial Training & Deployment Allowed) | ArXiv Standard Distribution; Permissive Open-Source Software (Apache-2.0) | 1 | Level 1: Plug-and-Play (Verified Package & Checkpoint Available) | 0 | 1 | Frank Li | 8 | [
"Frank Li",
"Bardia Khosravi",
"Mohammadreza Chavoshi",
"Theo Dapamede",
"YoungSeok Jeon",
"Janice Newsome",
"Hari Trivedi",
"Judy Gichoya"
] | [
"Clinical / Biomedical AI Research Institute"
] | http://arxiv.org/abs/2607.27235v1 | VERIFIED_LIVE | https://github.com/f10409/RadHarmony | [
"https://github.com/f10409/RadHarmony"
] | 11 | 3 | 2026-08-01 | Apache-2.0 | 60.58 | Training deep learning models on radiological images requires integrating heterogeneous datasets across different sources, file formats, directory layouts, label schemas, and annotation types. We present RadHarmony, an open-source Python library that provides a unified API for loading, harmonizing, and augmenting radio... | [
0.013389999978244305,
0.05670100077986717,
-0.03319000080227852,
-0.008934999816119671,
-0.06414400041103363,
-0.06582500040531158,
0.048833999782800674,
0.00381499994546175,
-0.039402998983860016,
0.058997999876737595,
0.001765000051818788,
-0.00394799979403615,
0.0984790027141571,
0.0563... | [
-0.031120000407099724,
0.011048000305891037,
-0.02872299961745739,
0.009955000132322311,
0.0215000007301569,
-0.07136300206184387,
-0.08515500277280807,
-0.013900999911129475,
-0.10819300264120102,
-0.09377399832010269,
-0.02870200015604496,
-0.060756999999284744,
0.06440400332212448,
0.05... | Medical AI, Clinical Foundation Models & Healthcare Intelligence | Local Open-Weights Clinical Model | 7B - 8B (Standard Clinical Model - BioMistral-7B/Llama-3) | 16 | 5.5 | Consumer GPU (RTX 4090 / 24GB) | [
"vLLM (High Throughput)",
"Ollama / llama.cpp (Local Clinic)",
"TGI",
"TensorRT-LLM"
] | [
"Medical Imaging & Multi-Center Diagnostic Radiography"
] | [
"f10409"
] | [
"Requires External Multi-Center Clinical Validation & Prospective Trial Testing"
] | git clone https://github.com/f10409/RadHarmony && cd RadHarmony && (pip install -e . || pip install -r requirements.txt) | Training deep learning models on radiological images requires integrating heterogeneous datasets across different sources, file formats, directory layouts, label schemas, and annotation types. | We present RadHarmony, an open-source Python library that provides a unified API for loading, harmonizing, and augmenting radiological datasets, with a primary focus on chest radiographs and early support for computed tomography (CT) and magnetic resonance imaging (MRI). | We demonstrate the library's utility by pretraining RadHarmony-ViT, a reference vision transformer baseline that combines three heterogeneous chest radiograph datasets with no dataset-specific code. | Explosive (>50/mo) | 192 | 2026-08-18T15:17:45.490494 |
2607.04688v1 | URSA: Chemistry-Aware Benchmark for Utilitarian Retrosynthesis Assessment | 2026-07-06T05:32:09Z | [
"cs.LG",
"cs.AI",
"cs.CE",
"cs.CL"
] | ArXiv Standard | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | true | 80 | Enterprise Safe (Commercial Training & Deployment Allowed) | ArXiv Standard Distribution | 1 | Level 1: Plug-and-Play (Verified Package & Checkpoint Available) | 0 | 1 | Bogdan Zagribelnyy | 10 | [
"Bogdan Zagribelnyy",
"Ivan Ilin",
"Nikita Bondarev",
"Anton Morgunov",
"Arkadii Lin",
"Maksim Kuznetsov",
"Rim Shayakhmetov",
"Vladimir Aladinskiy",
"Alex Aliper",
"Alex Zhavoronkov"
] | [
"Clinical / Biomedical AI Research Institute"
] | http://arxiv.org/abs/2607.04688v1 | VERIFIED_LIVE | https://github.com/insilicomedicine/URSA | [
"https://github.com/insilicomedicine/URSA"
] | 7 | 0 | 2026-08-03 | NOASSERTION | 56.34 | Synthesis planning aiming to find pathways of reactions for a target molecule is one of the most important and challenging tasks in drug discovery. Recent progress has produced both specialized deep-learning retrosynthesis systems and general-purpose large language models, but objective comparison remains difficult due... | [
-0.03812500089406967,
0.007544999942183495,
-0.04292700067162514,
-0.015886999666690826,
-0.07283700257539749,
-0.020809000357985497,
-0.09301900118589401,
0.05180000141263008,
-0.019940000027418137,
0.027867000550031662,
-0.011605000123381615,
-0.029218999668955803,
-0.06279099732637405,
... | [
-0.03274400159716606,
-0.05156400054693222,
-0.005288999993354082,
-0.026052000001072884,
-0.04181800037622452,
0.0010600000387057662,
-0.09837500005960464,
0.03425699844956398,
0.04124300181865692,
-0.009781000204384327,
-0.058065999299287796,
-0.0115769999101758,
-0.03208399936556816,
0.... | Medical AI, Clinical Foundation Models & Healthcare Intelligence | Clinical Decision Support & Algorithmic Pipeline | Model-Agnostic / Clinical Architecture | 0 | 0 | CPU Server / Model-Agnostic Clinical Host | [
"MONAI Medical Framework",
"Custom PyTorch Clinical Pipeline",
"FHIR / HL7 Adapter"
] | [
"Medical Imaging & Multi-Center Diagnostic Radiography"
] | [
"Quantitative Clinical Diagnostic & Statistical Evaluation"
] | [
"Requires External Multi-Center Clinical Validation & Prospective Trial Testing"
] | git clone https://github.com/insilicomedicine/URSA && cd URSA && (pip install -e . || pip install -r requirements.txt) | Synthesis planning aiming to find pathways of reactions for a target molecule is one of the most important and challenging tasks in drug discovery. | Recent progress has produced both specialized deep-learning retrosynthesis systems and general-purpose large language models, but objective comparison remains difficult due to the lack of flexible, chemically interpretable benchmarking protocols. | We find that while LLMs can support high-level strategic planning, they currently underperform specialized retrosynthesis models in reliably solving synthesis planning tasks. | Explosive (>50/mo) | 192 | 2026-08-18T15:17:53.609751 |
2607.09526v1 | ALICE: Learning a General-Purpose Pathology Foundation Model from Vision, Vision-Language, and Slide-Level Experts | 2026-07-10T15:35:06Z | [
"cs.CV",
"cs.AI"
] | ArXiv Standard | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | true | 60 | Internal R&D Only (Copyleft or Academic Terms) | ArXiv Standard Distribution; Repository License Unspecified | 2 | Level 2: Ready Codebase (Full Repository + Dependency Spec) | 0 | 1 | Jiawen Li | 8 | [
"Jiawen Li",
"Tian Guan",
"Huijuan Shi",
"Xitong Ling",
"Mingxi Fu",
"Anjia Han",
"Chao He",
"Yonghong He"
] | [
"Clinical / Biomedical AI Research Institute"
] | http://arxiv.org/abs/2607.09526v1 | VERIFIED_LIVE | https://github.com/WonderLandxD/ALICE | [
"https://github.com/WonderLandxD/ALICE"
] | 5 | 0 | 2026-07-17 | Unspecified | 54 | Foundation models are reshaping computational pathology, yet their capabilities remain shaped by pretraining objectives, data sources, and spatial scales, fragmenting complementary expertise across separate backbones. Here we present ALICE, a unified foundation model trained through multi-stage agglomerative distillati... | [
0.07175999879837036,
-0.07511100172996521,
0.025067999958992004,
-0.02143000066280365,
0.031415000557899475,
-0.02672399953007698,
0.019500000402331352,
0.04344300180673599,
-0.005758000072091818,
-0.03890100121498108,
-0.03410099819302559,
-0.032246001064777374,
0.04222799837589264,
0.113... | [
0.05525499954819679,
-0.13229599595069885,
0.030249999836087227,
-0.031209999695420265,
0.031612999737262726,
-0.07653500139713287,
-0.07287400215864182,
0.05150900036096573,
-0.002216000109910965,
-0.028432000428438187,
-0.10305599868297577,
0.006899000145494938,
0.03095800057053566,
0.11... | Medical AI, Clinical Foundation Models & Healthcare Intelligence | Clinical Decision Support & Algorithmic Pipeline | Model-Agnostic / Clinical Architecture | 0 | 0 | CPU Server / Model-Agnostic Clinical Host | [
"MONAI Medical Framework",
"Custom PyTorch Clinical Pipeline",
"FHIR / HL7 Adapter"
] | [
"Medical Imaging & Multi-Center Diagnostic Radiography"
] | [
"Quantitative Clinical Diagnostic & Statistical Evaluation"
] | [
"Requires External Multi-Center Clinical Validation & Prospective Trial Testing"
] | git clone https://github.com/WonderLandxD/ALICE && cd ALICE && (pip install -e . || pip install -r requirements.txt) | Foundation models are reshaping computational pathology, yet their capabilities remain shaped by pretraining objectives, data sources, and spatial scales, fragmenting complementary expertise across separate backbones. | Here we present ALICE, a unified foundation model trained through multi-stage agglomerative distillation that sequentially distills eight vision-only, vision-language, and slide-level teacher models into dedicated modules of a single backbone. | These results demonstrate that agglomerative distillation can consolidate complementary capabilities from specialized models into a unified backbone for broad computational pathology applications. | Explosive (>50/mo) | 192 | 2026-08-18T15:17:50.540061 |
2608.05615v1 | ALTER: Modeling Longitudinal Changes via Regional Differencing for 3D CT Report Generation | 2026-08-06T05:27:38Z | [
"cs.CV"
] | ArXiv Standard | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | true | 80 | Enterprise Safe (Commercial Training & Deployment Allowed) | ArXiv Standard Distribution; Permissive Open-Source Software (Apache-2.0) | 2 | Level 2: Ready Codebase (Full Repository + Dependency Spec) | 0 | 1 | Dongchen Li | 3 | [
"Dongchen Li",
"Jitao Liang",
"Wei Li"
] | [
"Clinical / Biomedical AI Research Institute"
] | http://arxiv.org/abs/2608.05615v1 | VERIFIED_LIVE | https://github.com/peytonkarlie/ALTER | [
"https://github.com/peytonkarlie/ALTER"
] | 4 | 0 | 2026-08-05 | Apache-2.0 | 53.5 | Computed tomography (CT) is widely used for clinical diagnosis and longitudinal follow-up, yet automatically generating accurate and complete radiology reports from three-dimensional (3D) CT remains challenging. Existing methods improve fine-grained correspondence between images and text by modeling anatomical regions,... | [
0.018869999796152115,
-0.038777001202106476,
0.09867499768733978,
0.006616000086069107,
-0.025157000869512558,
0.0107810003682971,
-0.14220400154590607,
-0.04941299930214882,
-0.03099299967288971,
-0.038385000079870224,
0.0035500000230968,
0.01902499981224537,
0.028397999703884125,
0.09450... | [
0.039772000163793564,
0.008813000284135342,
0.07113000005483627,
-0.02506300061941147,
0.007036000024527311,
0.001429000054486096,
-0.15938800573349,
-0.05443499982357025,
0.009511999785900116,
-0.07147400081157684,
-0.008125999942421913,
-0.030938999727368355,
0.02785000018775463,
0.14206... | Medical AI, Clinical Foundation Models & Healthcare Intelligence | Clinical Decision Support & Algorithmic Pipeline | Model-Agnostic / Clinical Architecture | 0 | 0 | CPU Server / Model-Agnostic Clinical Host | [
"MONAI Medical Framework",
"Custom PyTorch Clinical Pipeline",
"FHIR / HL7 Adapter"
] | [
"Medical Imaging & Multi-Center Diagnostic Radiography"
] | [
"Quantitative Clinical Diagnostic & Statistical Evaluation"
] | [
"Requires External Multi-Center Clinical Validation & Prospective Trial Testing"
] | git clone https://github.com/peytonkarlie/ALTER && cd ALTER && (pip install -e . || pip install -r requirements.txt) | Computed tomography (CT) is widely used for clinical diagnosis and longitudinal follow-up, yet automatically generating accurate and complete radiology reports from three-dimensional (3D) CT remains challenging. | We propose Anatomically Localized Temporal Evidence Representation (ALTER) to address these limitations. | ALTER achieves state-of-the-art results on most evaluation metrics across the RadGenome-ChestCT validation and CTRG-Chest-548K test sets. | Explosive (>50/mo) | 192 | 2026-08-18T15:17:40.637391 |
2608.09991v1 | Longitudinal 3D Foundation Modeling for Neoadjuvant Breast Cancer Response Prediction from Serial DCE-MRI | 2026-08-07T05:30:24Z | [
"eess.IV",
"cs.CV"
] | ArXiv Standard | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | true | 80 | Enterprise Safe (Commercial Training & Deployment Allowed) | ArXiv Standard Distribution; Permissive Open-Source Software (MIT) | 2 | Level 2: Ready Codebase (Full Repository + Dependency Spec) | 0 | 1 | Fidel Omar Tito Cruz | 6 | [
"Fidel Omar Tito Cruz",
"Neda Ghafouri",
"Zengyan Wang",
"Pegah Khosravi",
"Yu Tian",
"Chen Chen"
] | [
"Clinical / Biomedical AI Research Institute"
] | http://arxiv.org/abs/2608.09991v1 | VERIFIED_LIVE | https://github.com/omarftt/longitudinal_temporal_pillar | [
"https://github.com/omarftt/longitudinal_temporal_pillar"
] | 3 | 0 | 2026-08-02 | MIT | 51.6 | Pathologic complete response (pCR) is an important endpoint in neoadjuvant chemotherapy (NAC) for breast cancer, and predicting pCR from imaging during treatment could support treatment response assessment. Many existing imaging-based approaches rely on a single static timepoint, which fails to capture changes that occ... | [
-0.004327000118792057,
-0.12910200655460358,
0.033371999859809875,
-0.02460700087249279,
0.04466300085186958,
-0.052928000688552856,
-0.09787499904632568,
-0.0025100000202655792,
-0.0657379999756813,
-0.07209700345993042,
-0.10560700297355652,
-0.006535999942570925,
0.027153000235557556,
0... | [
-0.02913999930024147,
-0.11320699751377106,
0.036253999918699265,
-0.009860999882221222,
0.040839001536369324,
-0.031957000494003296,
-0.10973399877548218,
-0.023625999689102173,
0.016450000926852226,
-0.09886600077152252,
-0.08775900304317474,
0.015413999557495117,
-0.017090000212192535,
... | Medical AI, Clinical Foundation Models & Healthcare Intelligence | Clinical Decision Support & Algorithmic Pipeline | Model-Agnostic / Clinical Architecture | 0 | 0 | CPU Server / Model-Agnostic Clinical Host | [
"MONAI Medical Framework",
"Custom PyTorch Clinical Pipeline",
"FHIR / HL7 Adapter"
] | [
"Medical Imaging & Multi-Center Diagnostic Radiography"
] | [
"accuracy: 69.1%"
] | [
"Requires External Multi-Center Clinical Validation & Prospective Trial Testing"
] | git clone https://github.com/omarftt/longitudinal_temporal_pillar && cd longitudinal_temporal_pillar && (pip install -e . || pip install -r requirements.txt) | Pathologic complete response (pCR) is an important endpoint in neoadjuvant chemotherapy (NAC) for breast cancer, and predicting pCR from imaging during treatment could support treatment response assessment. | In this work, we present a longitudinal framework that combines a frozen 3D foundation encoder (Pillar-0) with our Temporal Dynamics Network (TDN) to predict treatment response from serial Dynamic Contrast-Enhanced (DCE) MRI acquired across four clinical timepoints from pre-treatment to pre-surgery. | Evaluated on 982 patients from the combined I-SPY2 and ACRIN-6698 cohort, the proposed model achieves strong performance across all reported metrics when longitudinal 3D imaging is fused with clinical data (test AUROC: 73.6%, balanced accuracy: 69.1%). | Explosive (>50/mo) | 192 | 2026-08-18T15:17:39.722873 |
2607.23794v1 | PathScale-R1: Cross-scale Reasoning for Pathological Image Analysis | 2026-07-26T18:36:23Z | [
"cs.CV",
"cs.AI"
] | ArXiv Standard | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | true | 60 | Internal R&D Only (Copyleft or Academic Terms) | ArXiv Standard Distribution; Repository License Unspecified | 2 | Level 2: Ready Codebase (Full Repository + Dependency Spec) | 0 | 1 | Chi Phan | 10 | [
"Chi Phan",
"Tianyi Zhang",
"Yufeng Wu",
"Qiaochu Xue",
"Jiajie Zhang",
"Linghan Cai",
"Zeyu Liu",
"Sudong Wang",
"Yueming Jin",
"Dan Hu"
] | [
"Clinical / Biomedical AI Research Institute"
] | http://arxiv.org/abs/2607.23794v1 | VERIFIED_LIVE | https://github.com/iMVR-PL/PathScale-R1 | [
"https://github.com/iMVR-PL/PathScale-R1"
] | 3 | 0 | 2026-08-11 | Unspecified | 51.12 | Pathological diagnosis is inherently multi-scale, requiring the integration of global tissue architecture at low magnification with cellular morphology at higher magnification. However, existing pathology benchmarks and vision-language models (VLMs) are still largely developed under single-scale settings, limiting thei... | [
0.06292600184679031,
-0.05447499826550484,
0.055991001427173615,
0.008175999857485294,
0.07767300307750702,
-0.0586169995367527,
0.003203999949619174,
0.0697460025548935,
-0.009440000168979168,
-0.013845999725162983,
-0.001243000035174191,
-0.06801000237464905,
0.008372999727725983,
0.2128... | [
0.05729899927973747,
-0.04880300164222717,
0.023207999765872955,
0.0370899997651577,
0.0628499984741211,
-0.029340000823140144,
0.00558600015938282,
0.0699940025806427,
-0.043584998697042465,
-0.013958999887108803,
-0.039326999336481094,
-0.11521299928426743,
0.009523999877274036,
0.154273... | Medical AI, Clinical Foundation Models & Healthcare Intelligence | Clinical Decision Support & Algorithmic Pipeline | Model-Agnostic / Clinical Architecture | 0 | 0 | CPU Server / Model-Agnostic Clinical Host | [
"MONAI Medical Framework",
"Custom PyTorch Clinical Pipeline",
"FHIR / HL7 Adapter"
] | [
"Medical Imaging & Multi-Center Diagnostic Radiography"
] | [
"Quantitative Clinical Diagnostic & Statistical Evaluation"
] | [
"Requires External Multi-Center Clinical Validation & Prospective Trial Testing"
] | git clone https://github.com/iMVR-PL/PathScale-R1 && cd PathScale-R1 && (pip install -e . || pip install -r requirements.txt) | Pathological diagnosis is inherently multi-scale, requiring the integration of global tissue architecture at low magnification with cellular morphology at higher magnification. | However, existing pathology benchmarks and vision-language models (VLMs) are still largely developed under single-scale settings, limiting their ability to learn clinically meaningful multi-magnification reasoning. | Extensive experiments demonstrate state-of-the-art performance of PathScale-R1 on cross-scale reasoning tasks and effective transfer to conventional single-scale pathology VQA. | Explosive (>50/mo) | 192 | 2026-08-18T15:17:45.363561 |
2608.08374v2 | Gated Spatial Redundancy Projection for Pathology Transformer Attentions | 2026-08-08T23:59:10Z | [
"cs.CV"
] | ArXiv Standard | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | true | 80 | Enterprise Safe (Commercial Training & Deployment Allowed) | ArXiv Standard Distribution | 2 | Level 2: Ready Codebase (Full Repository + Dependency Spec) | 0 | 1 | Zhiyuan Yang | 4 | [
"Zhiyuan Yang",
"Jiahao Cheng",
"Vincent Quoc-Huy Trinh",
"Mahdi S. Hosseini"
] | [
"Clinical / Biomedical AI Research Institute"
] | http://arxiv.org/abs/2608.08374v2 | VERIFIED_LIVE | https://github.com/AtlasAnalyticsLab/GatedSRP | [
"https://github.com/AtlasAnalyticsLab/GatedSRP"
] | 2 | 0 | 2026-08-11 | NOASSERTION | 49.14 | Transformer models are increasingly used for whole-slide image analysis in computational pathology. Yet, WSIs differ fundamentally from natural images: neighbouring patches often contain highly similar tissue type, stain, texture, and cellular composition. We identify this local spatial redundancy as a pathology-specif... | [
0.05305499956011772,
-0.06909099966287613,
-0.004904999863356352,
-0.08823099732398987,
-0.045545998960733414,
-0.02082899957895279,
0.008491000160574913,
0.06556099653244019,
0.034596998244524,
-0.011452999897301197,
-0.026186000555753708,
0.014635000377893448,
0.053011998534202576,
0.100... | [
0.028355000540614128,
-0.05967700108885765,
0.029745999723672867,
-0.012392000295221806,
0.0347369983792305,
-0.009003999643027782,
-0.030894000083208084,
0.05704599991440773,
0.05188800022006035,
-0.04948300123214722,
-0.032030001282691956,
-0.015184000134468079,
-0.020408999174833298,
0.... | Medical AI, Clinical Foundation Models & Healthcare Intelligence | Clinical Decision Support & Algorithmic Pipeline | Model-Agnostic / Clinical Architecture | 0 | 0 | CPU Server / Model-Agnostic Clinical Host | [
"MONAI Medical Framework",
"Custom PyTorch Clinical Pipeline",
"FHIR / HL7 Adapter"
] | [
"Medical Imaging & Multi-Center Diagnostic Radiography"
] | [
"Quantitative Clinical Diagnostic & Statistical Evaluation"
] | [
"Requires External Multi-Center Clinical Validation & Prospective Trial Testing"
] | git clone https://github.com/AtlasAnalyticsLab/GatedSRP && cd GatedSRP && (pip install -e . || pip install -r requirements.txt) | Transformer models are increasingly used for whole-slide image analysis in computational pathology. | We identify this local spatial redundancy as a pathology-specific failure mode of self-attention, where dominant neighbourhood features can be repeatedly mixed into patch-tokens and weaken subtle diagnostic or prognostic deviations. | Across five slide-level classification datasets, it improves the base attention on 12 of 16 reported metrics and achieves the best AUC on three datasets. | Explosive (>50/mo) | 192 | 2026-08-18T15:17:38.452270 |
2607.25108v1 | OPERA: Offline Policy-guided Expert Routing and Adaptation for Universal Biomedical Image Analysis | 2026-07-27T22:12:45Z | [
"cs.CV",
"cs.AI",
"cs.LG",
"eess.IV"
] | ArXiv Standard | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | true | 80 | Enterprise Safe (Commercial Training & Deployment Allowed) | ArXiv Standard Distribution; Permissive Open-Source Software (MIT) | 1 | Level 1: Plug-and-Play (Verified Package & Checkpoint Available) | 0 | 1 | Zihan Li | 5 | [
"Zihan Li",
"Feiyang Liu",
"Dandan Shan",
"Ruibo Wang",
"Qingqi Hong"
] | [
"Clinical / Biomedical AI Research Institute"
] | http://arxiv.org/abs/2607.25108v1 | VERIFIED_LIVE | https://github.com/HUANGLIZI/OPERA | [
"https://github.com/HUANGLIZI/OPERA"
] | 2 | 0 | 2026-07-30 | MIT | 48.66 | Biomedical image analysis spans diverse modalities and tasks, yet real-world deployment is hindered by severe distribution shifts across scanners, protocols, and patient populations. High-performing models consequently require repeated domain-specific fine-tuning, which is a costly cycle that becomes impractical when l... | [
0.008918000385165215,
-0.031380001455545425,
0.02560500055551529,
0.0009130000253207982,
0.044245000928640366,
-0.07990600168704987,
-0.07564099878072739,
0.04486599937081337,
-0.05651099979877472,
-0.013659000396728516,
-0.07818800210952759,
0.0018810000037774444,
-0.017020000144839287,
0... | [
-0.01715799979865551,
-0.05894099920988083,
-0.008640999905765057,
0.0517750009894371,
0.04339300096035004,
-0.09447000175714493,
-0.026861000806093216,
-0.0006649999995715916,
-0.0494999997317791,
-0.06366799771785736,
-0.07535500079393387,
-0.024675000458955765,
0.04098600149154663,
0.06... | Medical AI, Clinical Foundation Models & Healthcare Intelligence | Local Open-Weights Clinical Model | 7B - 8B (Standard Clinical Model - BioMistral-7B/Llama-3) | 16 | 5.5 | Consumer GPU (RTX 4090 / 24GB) | [
"vLLM (High Throughput)",
"Ollama / llama.cpp (Local Clinic)",
"TGI",
"TensorRT-LLM"
] | [
"Medical Imaging & Multi-Center Diagnostic Radiography"
] | [
"Quantitative Clinical Diagnostic & Statistical Evaluation"
] | [
"Distribution Shift & Multi-Center Generalization",
"Data Privacy & HIPAA / GDPR Boundaries"
] | git clone https://github.com/HUANGLIZI/OPERA && cd OPERA && (pip install -e . || pip install -r requirements.txt) | Biomedical image analysis spans diverse modalities and tasks, yet real-world deployment is hindered by severe distribution shifts across scanners, protocols, and patient populations. | We propose OPERA (Offline Policy-guided Expert Routing and Adaptation), a multi-agent ensemble framework that addresses this deployment bottleneck by treating expert weight assignment as an offline policy learning problem: a routing policy is learned from a small validation set without gradient updates to any expert ag... | OPERA consistently improves performance and calibration quality, demonstrating that offline policy-guided expert agents coordination is a practical path to deployable biomedical AI without retraining. | Explosive (>50/mo) | 192 | 2026-08-18T15:17:45.193409 |
2607.11257v1 | LaGuadia: Language-Guided Adaptive Distillation from Pathology Foundation Models | 2026-07-13T08:38:21Z | [
"cs.CV",
"cs.LG"
] | ArXiv Standard | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | true | 80 | Enterprise Safe (Commercial Training & Deployment Allowed) | ArXiv Standard Distribution; Permissive Open-Source Software (Apache-2.0) | 2 | Level 2: Ready Codebase (Full Repository + Dependency Spec) | 0 | 1 | Gangsu Kim | 2 | [
"Gangsu Kim",
"Won-Ki Jeong"
] | [
"Clinical / Biomedical AI Research Institute"
] | http://arxiv.org/abs/2607.11257v1 | VERIFIED_LIVE | https://github.com/hvcl/LaGuadia | [
"https://github.com/hvcl/LaGuadia"
] | 2 | 0 | 2026-07-14 | Apache-2.0 | 48.1 | "Pathology Foundation Models (PFMs) offer powerful Whole Slide Image (WSI) representations but suffe(...TRUNCATED) | [0.03434399887919426,-0.13674800097942352,0.08779700100421906,-0.02423500083386898,-0.01293599978089(...TRUNCATED) | [0.03911000117659569,-0.07696600258350372,0.02790600061416626,-0.030685000121593475,-0.0020290000829(...TRUNCATED) | Medical AI, Clinical Foundation Models & Healthcare Intelligence | Clinical Decision Support & Algorithmic Pipeline | Model-Agnostic / Clinical Architecture | 0 | 0 | CPU Server / Model-Agnostic Clinical Host | [
"MONAI Medical Framework",
"Custom PyTorch Clinical Pipeline",
"FHIR / HL7 Adapter"
] | [
"Medical Imaging & Multi-Center Diagnostic Radiography"
] | [
"Quantitative Clinical Diagnostic & Statistical Evaluation"
] | [
"Requires External Multi-Center Clinical Validation & Prospective Trial Testing"
] | "git clone https://github.com/hvcl/LaGuadia && cd LaGuadia && (pip install -e . || pip install -r re(...TRUNCATED) | "Pathology Foundation Models (PFMs) offer powerful Whole Slide Image (WSI) representations but suffe(...TRUNCATED) | "We propose LaGuadia (Language-Guided Adaptive DistillAtion), a framework that develops a compact pa(...TRUNCATED) | "These results highlight clinical language as an effective semantic anchor for building efficient an(...TRUNCATED) | Explosive (>50/mo) | 192 | 2026-08-18T15:17:49.897514 |
End of preview. Expand in Data Studio
🏥 Medical AI & Clinical Foundation Models Dataset (2026 Edition)
Sample dataset of 30 audit-verified research papers covering Clinical LLMs, Medical Foundation Models, Radiology/Pathology Vision, and EHR Intelligence with 384d PyTorch embeddings.
🛒 Full 1,000 Paper B2B Dataset Available on Gumroad
Get the complete 1,000 paper dataset (1,000 papers + OpenAlex Citations + IP Safety Score + SQLite/CSV/Parquet + Quickstart Script) on Gumroad: 👉 Get Full 1,000 Dataset on Gumroad ($19 / $39 / $89)
- Downloads last month
- 35