Tag
Introduces FinInvest-GTCN, a graph-temporal-causal network for risk-aware venture capital investment decisions, achieving state-of-the-art risk-adjusted returns with explainable predictions.
AnTenA is an interpretable system that leverages large language models to generate natural language explanations for hidden patterns extracted via tensor decomposition, without relying on potentially inaccurate labels or metadata.
This paper presents a multi-stage explainable framework that combines SHAP-based token attribution, theory-informed linguistic features, and LLaMA-3.1-70B-Instruct LLM reasoning to interpret transformer-based speech models for cognitive impairment detection, achieving strong clinical alignment and high usability scores.
This paper proposes KARMA, a method for explaining multivariate time series forecasting models by constructing a K-order Markov surrogate model that captures temporal dependencies, offering a five-level global explanation hierarchy.
This paper applies ensemble machine learning models (Random Forest, Gradient Boosting, XGBoost, Extra Trees) to detect cirrhosis in hepatitis C patients using 28 features from 2038 Egyptian patients. The Extra Trees model achieved 96.92% accuracy with only 16 features, outperforming other models.
This paper proposes a confusion matrix-based graph construction method and a hybrid loss function for Graph Neural Networks to improve multi-site pollution prediction accuracy and interpretability, evaluated on real-world air pollution data.
The paper proposes using spectral entropy as a metric to quantify noise introduced by explainability techniques in ECG arrhythmia classification, helping to distinguish true model signal from XAI-generated artifacts.
IONS is an open-source approach to AI memory and reasoning that uses a graph of evidence-backed claims called Cognitive Building Blocks (CBBs) to store knowledge outside model weights, making reasoning inspectable.
This paper proposes a hybrid predictive model combining ensemble feature selection (ANOVA and mutual information) with Harris Hawks optimization-tuned logistic regression for explainable mental health risk prediction in female sex workers, achieving 95.78% accuracy.
This paper presents a scalable framework using LLMs for implicit sentiment analysis of product desirability from qualitative feedback, achieving up to 0.97 Pearson correlation and 94% accuracy while providing explanations, with GPT-4o-mini offering similar performance at 94% lower cost.
TelcoAgent is a foundation model-based framework for scalable and explainable multi-KPM forecasting in 5G networks, using automated 3GPP knowledge graph construction and a time-series foundation model for zero-shot prediction.
Introduces Evolving Programmatic Bottlenecks (EPB), a framework for interpreting neural combinatorial optimization policies by distilling black-box models into human-readable program portfolios using LLM-guided evolution.
GLARE is an LLM-based interface that translates natural language questions into SQL queries over local explanation data, enabling users to interactively explore global explanations of black-box image classifiers.
The PRAG framework combines traditional RAG with a Paninian rule engine for safer medical AI, achieving a 71% reduction in unsafe answers on MedQA. It provides auditable rule traces and is open-sourced.
Proposes a reinforcement learning-based post-training method using Group Relative Policy Optimization (GRPO) and chain-of-thought supervision to improve classification and explanation quality for hateful and propagandistic meme detection in thinking-based multimodal large language models, achieving improvements on the Hateful Memes and ArMeme benchmarks.
A study presenting a cross-method explainability audit of the BridgeDPI drug-target interaction model, combining gradient-based attributions and occlusion to reveal modality dominance and artifacts, providing testable hypotheses for drug discovery.
ProtoX-AD is a prototype-based self-explainable framework for self-supervised time series anomaly detection that provides interpretable explanations for detected anomalies by learning transformation-aware prototypes, achieving performance comparable to black-box methods while offering semantic anomaly characterization.
This paper proposes a lightweight monitoring primitive for tracking interpretable parameter–KPI dependencies in AI-integrated Radio Access Networks (AI-RAN) to enable conflict detection and slow-loop model refresh, using Boolean matrix representations and sliding-window inference.
This paper proposes a Glassbox Framework that uses Bayesian networks as transparent ante-hoc mediation layers for generative models, enabling auditable reasoning traces and contestable outputs to address opacity in high-stakes AI applications.
TRIAGE is a framework that trains LLMs to generate dialectical reasoning for continuous risk scoring from irregularly sampled medical time series, achieving improved calibration and interpretability.