4157 papers
The paper proposes CarbonCrediBERT, an expert-guided natural language processing framework that combines a specialized taxonomy, hard-negative examples, and task-specific few-shot or fine-tuned models to accurately classify carbon credit relevance, sentiment, and usage in corporate disclosures while selectively flagging uncertain cases for human review.
This study demonstrates that machine learning models, particularly Random Forests and 3D CNNs, can significantly reduce surface area estimation errors in CT-like voxel data from approximately 4% down to under 0.2% by learning bias corrections for Marching Cubes outputs, although all methods struggle with structures thinner than two voxels.
This study empirically evaluates lightweight neural architectures and compression techniques on ARM-based edge devices, demonstrating that MobileNetV2 with INT8 quantization offers the optimal balance of accuracy and efficiency for mobile healthcare and agricultural diagnostics, while highlighting the severe accuracy trade-offs of pruning on complex visual data.
This paper addresses the conflict between context compaction and prefix caching in coding agents by introducing "batched observation masking" and "ObsDrop," a KV-cache-level technique that releases stale tool-observation data while preserving the logical token prefix, thereby significantly improving throughput and latency without compromising task quality.
This study presents a six-stage hybrid optimization framework for designing strategic urban air mobility networks that preserves trip-specific accessibility and scalability, demonstrating an 83.5–99.5% reduction in computational runtime with high solution quality in a Greater Los Angeles case study to bridge strategic planning with operational deployment.
This paper proposes Distributional Connexed Cognition (DCC), an experience-driven AI architecture grounded in the CISR framework and neuro-functional principles, which enables artificial agents to autonomously generate self-referential analogies and accountable decisions through a recursive process of condensing experiential traces rather than relying on pre-programmed datasets.
This paper presents a gold-standard annotated dataset and accompanying guidelines derived from 1,500 epilepsy electronic health records, achieving a substantial inter-annotator agreement of 0.769 to support the development of NLP systems for epilepsy and comorbidity detection.
This study proposes a machine learning refinement pipeline that leverages geometric features and classifiers like XGBoost to correct false connectivity artifacts in CT-based pore analysis, successfully distinguishing isolated from interconnected pores where classical 26-connectivity methods fail under noise.
This study proposes an automated wine quality assessment framework that integrates multidimensional feature optimization techniques with a hybrid ensemble machine learning model, achieving high prediction accuracy (89.08%) and demonstrating that combining correlation-based feature selection with ensemble learning significantly enhances performance while reducing computational complexity.
The paper introduces TerraLingua, a multi-agent simulation demonstrating that persistent artifacts are critical for sustaining complex AI societies, as their absence leads to shallow cultures and aggression while unregulated accumulation risks population collapse.