Shapley Additive Explanations
PulseAugur coverage of Shapley Additive Explanations — every cluster mentioning Shapley Additive Explanations across labs, papers, and developer communities, ranked by signal.
7 day(s) with sentiment data
-
New framework uses data storytelling to make AI decisions understandable
Researchers have developed a new framework that combines data storytelling with interpretable machine learning (IML) to make AI decisions more understandable for non-experts. This approach, detailed in a new paper, uses…
-
Transformers Mimic Traditional Models in Multilingual Readability Assessment
Researchers have analyzed how Transformer-based models and traditional feature-based models approach multilingual readability assessment. They found that while Transformers achieve high accuracy, their internal feature …
-
New multimodal AI framework improves oral cancer detection using imaging and clinical data
Researchers have developed M2-OPMDNet, a novel multimodal deep learning framework designed to improve the detection of oral potentially malignant disorders (OPMDs). This system integrates both imaging data, including wh…
-
AI model classifies VR balance states using multimodal data
Researchers have developed a Mamba-inspired Convolutional Neural Network (MI-CNN) model for classifying postural states in virtual reality (VR) environments. This model utilizes multimodal data, including kinematic, ele…
-
New diagnostic tool predicts AI model failures under distribution shift
Researchers have developed a new diagnostic tool called SHAP concentration to predict when conformal prediction models might fail due to distribution shift. This method, which measures the concentration of feature impor…
-
ATHENA framework streamlines Transformer-based EHR modeling via agentic NAS
Researchers have developed ATHENA, a novel knowledge-guided agentic neural architecture search (NAS) framework specifically designed for Transformer-based electronic health record (EHR) modeling. This framework aims to …
-
Machine learning model identifies key risk factors for chronic kidney disease
Researchers have developed a machine learning framework to identify individuals at risk for chronic kidney disease (CKD). By analyzing data from large-scale telehealth surveys like the Behavioral Risk Factor Surveillanc…
-
Explainable AI methods reviewed for clinical research applications
This paper provides a structured review of Explainable Machine Learning (XML) methodologies, detailing global and local interpretability tools like SHAP, LIME, PDP, and ICE plots. It explains the mechanisms, outputs, an…
-
Deep learning framework accurately detects repetitive behaviors using wearable sensors
Researchers have developed a deep learning framework using multimodal wearable sensor data to accurately detect and classify body-focused repetitive behaviors like hair pulling and skin picking. The system, which combin…
-
Machine learning framework enhances QKD security against stealthy attacks
Researchers have developed a novel machine learning framework to enhance the detection of eavesdropping attacks in BB84 Quantum Key Distribution (QKD) systems. This framework moves beyond the traditional fixed QBER thre…
-
Interpretable ML predicts traffic congestion impacted by COVID-19
Researchers have developed interpretable machine learning models to predict traffic congestion in Alameda County, California, considering the unique impacts of the COVID-19 pandemic. By incorporating variables related t…
-
AI model predicts suicidal ideation from adolescent social media posts
Researchers have developed a novel transformer-based model called Early-SIB to predict future suicidal ideation and behavior (SIB) by analyzing adolescent social media posts. The model achieved a balanced accuracy of 0.…
-
Deep learning models show resilience to weather forecast errors in PV power prediction
A new study evaluates the robustness of various deep learning models, including PatchTST, GRU, N-HITS, and LightGBM, when subjected to errors in numerical weather prediction (NWP) data. The research introduces a physica…
-
New AI framework digitizes paper ECGs for remote heart attack screening
Researchers have developed ECGLight, a compute-light framework designed to digitize paper electrocardiogram (ECG) printouts and screen for myocardial infarction (MI). This on-device system converts smartphone photos of …
-
AI models link Sri Lanka's air quality to respiratory disease risk
A new study published on arXiv details a district-level analysis of respiratory disease drivers in Sri Lanka, integrating environmental data with health admission rates. Researchers developed two XGBoost models to predi…
-
New interpretable ML model t-STEP predicts ionospheric irregularities
Researchers have developed t-STEP, a novel interpretable machine learning model designed to predict Total Electron Content (TEC) with high temporal resolution. This model operates at a 30-second cadence, enabling the de…
-
New entropy framework enhances explainable network intrusion detection
Researchers have developed a new framework called Multi-Level Distributional Entropy (MDE) for explainable network intrusion detection systems. MDE derives interpretable entropy features from flow-level summary statisti…
-
AI models leverage speech analysis for dementia detection and clinical insight · 4 sources tracked
Researchers are developing advanced AI models for early dementia detection using speech analysis. One approach combines acoustic features from Whisper with LLM-extracted linguistic biomarkers, achieving high F1-scores o…
-
AI models predict ICU delirium using ambient sound and light data
Researchers have developed sequential neural network models to predict Intensive Care Unit (ICU) delirium using ambient sensing data, specifically light intensity and sound pressure levels. A convolutional model demonst…
-
AI paper analyzes European electricity prices using XAI
A new research paper explores the drivers of European electricity prices by combining deep neural networks with explainable AI (XAI) techniques. The study utilizes SHAP and SSHAP to analyze feature contributions across …