View PDF ↗
PDF Viewer
Open Full PDF

BUILDER'S SANDBOX

Build This Paper

Use an AI coding agent to implement this research.

OpenAI Codex
OpenAI CodexAI Agent

Lightweight coding agent in your terminal.

Claude Code
Claude CodeAI Agent

Agentic coding tool for terminal workflows.

AntiGravity IDE
AntiGravity IDEScaffolding

AI agent mindset installer and workflow scaffolder.

Cursor
CursorIDE

AI-first code editor built on VS Code.

VS Code
VS CodeIDE

Free, open-source editor by Microsoft.

MVP Investment

$10K - $13K
6-10 weeks
Engineering
$8,000
Cloud Hosting
$240
SaaS Stack
$800
Domain & Legal
$500

6mo ROI

2-4x

3yr ROI

10-20x

Lightweight AI tools can reach profitability quickly. At $500/mo average contract, 20 customers = $10K MRR by 6mo, 200+ by 3yr.

Talent Scout

S

Samah Fodeh

Yale School of Medicine

L

Linhai Ma

Yale School of Medicine

Y

Yan Wang

Yale School of Medicine

S

Srivani Talakokkul

Yale School of Medicine

Find Similar Experts

Healthcare experts on LinkedIn & GitHub

References (68)

[1]
A general approach to improve adversarial robustness of DNNs for medical image segmentation and detection
2024Linhai Ma, Jiasong Chen et al.
[2]
Improving adversarial robustness of deep neural networks via adaptive margin evolution
2023Linhai Ma, Liang Liang
[3]
Identifying the mechanisms of patient-centred communication in secure messages between clinicians and cancer patients
2023Aantaki Raisa, J. Alpert et al.
[4]
Towards lifting the trade-off between accuracy and adversarial robustness of deep neural networks with application on COVID 19 CT image classification and medical image segmentation
2023Linhai Ma, Liang Liang
[5]
Extracting social determinants of health events with transformer-based multitask, multilabel named entity recognition
2023Russell Richie, V. Ruiz et al.
[6]
The 2022 n2c2/UW Shared Task on Extracting Social Determinants of Health
2023Kevin Lybarger, Meliha Yetisgen-Yildiz et al.
[7]
TwHIN-BERT: A Socially-Enriched Pre-trained Language Model for Multilingual Tweet Representations at Twitter
2022Xinyang Zhang, Yury Malkov et al.
[8]
Few-Shot Learning for Clinical Natural Language Processing Using Siamese Neural Networks
2022David Oniani, Sonish Sivarajkumar et al.
[9]
Adaptive Adversarial Training to Improve Adversarial Robustness of DNNs for Medical Image Segmentation and Detection
2022Linhai Ma, Liang Liang
[10]
Dependency multi-weight-view graphs for event detection with label co-occurrence
2022Yan Wang, Jian Wang et al.
[11]
Conditional Probability Joint Extraction of Nested Biomedical Events: Design of a Unified Extraction Framework Based on Neural Networks
2022Yan Wang, Jian Wang et al.
[12]
Training language models to follow instructions with human feedback
2022Long Ouyang, Jeff Wu et al.
[13]
A Regularization Method to Improve Adversarial Robustness of Neural Networks for ECG Signal Classification
2021Linhai Ma, Liang Liang
[14]
Patient-generated health data and electronic health record integration: a scoping review
2020V. Tiase, William Hull et al.
[15]
Improving secure messaging: A framework for support, partnership & information-giving communicating electronically (SPICE).
2020J. Alpert, Shu Wang et al.
[16]
A Review of Generalized Zero-Shot Learning Methods
2020Farhad Pourpanah, Moloud Abdar et al.
[17]
Self-Alignment Pretraining for Biomedical Entity Representations
2020Fangyu Liu, Ehsan Shareghi et al.
[18]
Using de-identified electronic health records to research mental health supported housing services: A feasibility study
2020C. Dalton-Locke, J. Thygesen et al.
[19]
Increasing-Margin Adversarial (IMA) Training to Improve Adversarial Robustness of Neural Networks
2020Linhai Ma, Liang Liang
[20]
Identifying Predictors of Suicide in Severe Mental Illness: A Feasibility Study of a Clinical Prediction Rule (Oxford Mental Illness and Suicide Tool or OxMIS)
2020Morwenna Senior, M. Burghart et al.

Showing 20 of 68 references

Founder's Pitch

"PVminer efficiently identifies and analyzes patient voices in healthcare data for enhanced patient-centered care."

Healthcare NLPScore: 7View PDF ↗

Commercial Viability Breakdown

0-10 scale

High Potential

3/4 signals

7.5

Quick Build

4/4 signals

10

Series A Potential

2/4 signals

5

Sources used for this analysis

arXiv Paper

Full-text PDF analysis of the research paper

GitHub Repository

Code availability, stars, and contributor activity

Citation Network

Semantic Scholar citations and co-citation patterns

Community Predictions

Crowd-sourced unicorn probability assessments

Analysis model: GPT-4o · Last scored: 2/24/2026

🔭 Research Neighborhood

Generating constellation...

~3-8 seconds

Why It Matters

Understanding the patient voice using tools like PVminer allows healthcare providers to better tailor their approaches to meet individual patient needs, reinforcing patient-centered care, improving outcomes, and reducing healthcare costs.

Product Angle

PVminer can be productized as a software tool for healthcare systems to analyze patient-generated data, aiding in patient-centric analysis and care improvement strategies.

Disruption

PVminer could replace manual, labor-intensive qualitative analysis of patient communications, offering scalable and precise automated insights aligned with patient and provider needs.

Product Opportunity

The healthcare sector faces challenges in patient communication and data management. Tools that extract meaningful insights from patient data are in demand for improving care outcomes, with hospitals and clinics as primary buyers.

Use Case Idea

Develop an API enabling healthcare providers to integrate patient voice analysis into existing e-health solutions, aiding improved patient-provider communication and decision-making processes.

Science

PVminer uses a patient-tailored BERT model for detecting and categorizing patient voices in text data, including surveys and secure messages, splitting the task into label prediction for structured representation of both communicative and social aspects.

Method & Eval

The tool was evaluated using pre-trained, patient-specific BERT models fine-tuned on a dataset of patient-generated messages. It demonstrated superior performance over existing models for predicting communications and social determinants.

Caveats

Potential limitations include model biases due to demographic variations, and the need for regular updates to handle evolving linguistic preferences and new data types.

Author Intelligence

Samah Fodeh

LEAD
Yale School of Medicine
samah.fodeh@yale.edu

Linhai Ma

Yale School of Medicine

Yan Wang

Yale School of Medicine

Srivani Talakokkul

Yale School of Medicine

Ganesh Puthiaraju

Yale School of Medicine

Afshan Khan

Yale School of Medicine

Ashley Hagaman

Yale School of Public Health

Sarah Lowe

Yale School of Public Health

Aimee Roundtree

Texas State University