Hackathons: Kaggle (page 3)
Hackathons · Kaggle: 157 entries in total; this page lists entries 31–45.
HedgeDecode
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Benchmarking social cognition through pragmatic policy inference in hedged conversation
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
AdvisorBench: Do AI Models Widen the Advisory Divide?
Evidence grade: for discovery only; must not by itself support a formal conclusion.
A social cognition benchmark for testing whether AI advisors give worse guidance to users who communicate at lower literacy levels.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
MEDLEY-BENCH: Behavioral Metacognition Under Social Pressure
Evidence grade: for discovery only; must not by itself support a formal conclusion.
The first benchmark to mathematically separate self-revision from social conformity in LLMs
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
LearningBench: Measuring Inference-Time Learning in LLMs
Evidence grade: for discovery only; must not by itself support a formal conclusion.
6 learning sub-abilities, 135 novel tasks - scoring not just whether a model learns something it has never seen, but how efficiently it does
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
GAUGE: Measuring the gap between what LLMs know and what they do about it
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Do LLMs act on their own uncertainty? 3-turn metacognition benchmark. 8 models, 4 families. Gemini never abstains. Claude and Qwen do.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
metaproteus
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Metacognitive Self-Knowledge in Language Models
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
EpiCast: AI-Powered Syndromic Surveillance for ECOWAS West Africa
Evidence grade: for discovery only; must not by itself support a formal conclusion.
AI-powered syndromic surveillance for West Africa. On-device clinical extraction in 6 languages. No internet required. Built for CHWs.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Sunny - Private Skin Health Tracker
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Sunny is an iOS application which runs a fine-tuned version of MedGemma on-device for private skin health tracking.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
FieldScreen AI
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Offline Multi-Model TB Screening for Resource-Limited Settings
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Tracer: AI-Powered Diagnostic Loop Tracking
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Fine-tuned MedGemma that closes the loop between test orders and results before patients fall through the cracks.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
CaseTwin: Clinical Case Matching & Referral Powered by MedGemma
Evidence grade: for discovery only; must not by itself support a formal conclusion.
MedGemma-powered clinical decision support: match chest X-rays to historical twins and route patients to the right specialist in minutes.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
BigTB6
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Multimodal Voice Driven Preliminary TB Screening System
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
ClinicDx: Clinical AI for Africa's Under-Resourced Health Centers
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Extending MedGemma 4B with medical voice and knowledge grounded reasoning to deliver CDS, scribe, imaging, and OCR natively in OpenMRS.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
UniRad3s: Reimagining Radiology with Spot, Segment and Simplify
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Redefining radiology workflow through fast and accurate automated lesion detection, segmentation, and reporting.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
BridgeDx: AI-Powered Clinical Decision Support for Frontline Health Workers
Evidence grade: for discovery only; must not by itself support a formal conclusion.
BridgeDx is a fully offline clinical decision support tool that runs MedGemma locally via Ollama to deliver protocol-grounded differential d
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24