2661 papers
This study demonstrates that enforcing explicit evidence-admission requirements in clinical question answering significantly alters route rankings and latency metrics while reducing material evidence-validity errors, highlighting the need for standardized admission rules in clinical latency benchmarks.
This paper proposes a two-layer framework featuring candidate evidence reranking (CAPE) and risk-aware answer selection (CTA/EBC) to optimize evidence allocation and answer choice in multi-hop retrieval-augmented generation, significantly improving recall and F1 scores over existing baselines like SAG, RRF, and majority voting.
This pilot study demonstrates that small, open-weight language models applied to psychiatric knowledge questions exhibit poor confidence calibration, domain-dependent accuracy rankings, and that prompting them to abstain from answering often degrades rather than improves performance when evaluated with rigorous paired statistics.
This study introduces the Collective Epistemic Network Framework (CENF), a computational model integrating adaptive field control and trust plasticity that demonstrates significant improvements in consensus accuracy and stability against adversarial attacks in synthetic multi-agent networks, while establishing formal convergence guarantees under specific design conditions.
This paper presents a six-axis survey and controlled benchmark demonstrating that while deep neural networks like R3D-18, R(2+1)D-18, and MC3-18 achieve high accuracy in human action recognition, practical deployment requires balancing recognition performance with computational efficiency, temporal robustness, and resource constraints rather than relying on accuracy alone.
This study demonstrates that a frozen pretrained VideoMAE encoder combined with a lightweight linear classifier enables highly label-efficient human action recognition, achieving strong performance on UCF101 and HMDB51 benchmarks with minimal annotations while maintaining stable optimization and constant computational resource usage.
This study develops a machine learning system using a Thales manufacturing dataset to predict three-class manufacturing efficiency, achieving up to 100% accuracy with a Gradient Boosting classifier and deploying the solution via an interactive Streamlit dashboard.
This pilot study demonstrates that for small open-weight language models (1.2B–9.2B parameters) applied to medical and psychiatric tasks, extracting internal token probabilities is a significantly more reliable method for error detection than relying on the models' verbally expressed confidence, particularly for the smallest models where verbalized confidence performs at chance levels.
This paper introduces HRT-LI, a certified dynamic learned index for hierarchical string keys that maintains strict rank error guarantees by coupling a frozen predictive model with a ledger-based correction mechanism, validated through extensive experiments on hundreds of millions of real-world strings.
This synthetic-vignette study demonstrates that while adding a safety-priming instruction to small open-weight language models can significantly reduce the omission of critical risk factors in psychiatric note summarization for some models, it often creates a trade-off by increasing the omission of routine clinical details, suggesting that such safety interventions must be validated on a per-model basis rather than assumed to be universally effective.