eva-bench grows to three enterprise voice agent domains
eva-bench data 2.0 adds itsm and healthcare hrsd domains, totaling 213 scenarios across 121 tools to test voice agents on realistic, domain-specific tasks.
aitag research
summaries filed under research
ai emotional dependence grows from routine use, not just chatbots
new research finds emotional reliance on ai often starts incidentally during everyday tasks, not through dedicated companion apps, and can shift preferences away from human support.
aipre-deployment ai agent verification with ontology and trust certificates
a framework uses ontologies to generate test scenarios and issue verifiable trust certificates for enterprise ai agents before deployment.
aifinite-iteration theory for tensor power iteration
new analysis of alternating power iteration for spiked tensor models gives finite-iteration error bounds and explains warm-start behavior without relying on specific initializations.
aiieee p3109 arithmetic formats for machine learning
the ieee p3109 draft standard defines parameterized binary floating-point formats and operations to support efficient, consistent machine learning computations.
aipeel framework exposes ai distortions in research
a semiotic scaffolding called peel reveals systematic distortions in ai-generated text condensations, showing fluency does not equal fidelity.
aieeg region contributions to cognitive workload prediction
a large-scale analysis across four eeg datasets examines how different scalp regions contribute to predicting cognitive workload, revealing consistent patterns and practical implications for sensor selection.
aifive papers that explain llms clearly
a guide to five foundational papers covering transformer architecture, few-shot learning, scaling laws, instruction tuning, and retrieval-augmented generation for understanding large language models.
aibehaviorbench tests ai on real user decisions
a new benchmark uses public prediction-market and blockchain data to evaluate how well models predict individual beliefs and trades.
aigoogle open-sources hydrology model for flood forecasting
google research releases its ai flood forecasting framework on github, enabling national agencies to train and customize models with local data.
aidpo cuts text degeneration in ocr models
direct preference optimization reduced text degeneration by an average of 59.4% across five ocr model families by using the model's own failure outputs as rejection pairs.
aihuman-gated bandits speed up rental pricing
a new framework lets human agents approve or reject algorithmic price suggestions, using old pricing data to skip the slow start typical in sparse booking markets.
aitera cuts derivative gp cost to linear scaling
a new method reduces the cubic complexity of gaussian processes with gradients by using exact gradient reduction and vecchia approximation.
aistructure over models for scientific discovery
a position paper argues that in high-dimensional settings, many different mechanisms can produce the same data, so predictive success does not prove a model has found the true mechanism, and large language models can hide this by giving a single fluent explanation.
aiaction-gated memory keeps robot policies at constant vram
aura-mem uses a learned gate to write only when observations change actions, keeping memory fixed at 4,224 bytes regardless of episode length.
aitarget updates may stabilize linear q-learning
periodic and soft target updates can guarantee convergence in linear q-learning under explicit spectral and step-size conditions.
aigradient-based early stopping for boosted trees
a new method uses gradient tests instead of validation loss to decide when to stop training gradient boosted trees, avoiding the need for a patience parameter.
aiclass-split anomaly detection can fail when classes overlap
a new diagnostic reveals that common anomaly detection benchmarks become unreliable when held-out classes overlap with normal data in representation space.
aialigning ehr models with llms for clinical reasoning
a new framework aligns structured electronic health record representations with large language models to improve clinical prediction and reasoning.
aimicrosoft releases mai models with low active parameters
microsoft announced two new language models, mai-thinking-1 and mai-code-1-flash, with low active parameter counts and claims of clean training data.
ai