LIVE PULSE
4.0 Anthropic CEO Amodei calls for slower AI development and shared safety rules — 11 src2.2 Agility Robotics unveils Digit 5 humanoid for warehouses and factories — 2 src2.0 Apple ships rebuilt Siri with Google Gemini, but not in the EU — 2 src1.8 Siri AI in macOS 27 Golden Gate: FAQ, Germany availability, privacy questions — 2 src1.4 Sam Altman says OpenAI will not go public in 2026, citing AI safety concerns — 5 src1.1 OpenAI contractors review real ChatGPT conversations to rate responses, report says — 2 src1.1 Anthropic data retention policy prompts firms to limit Claude use for sensitive work — 1 src1.1 Study examines issue bias in LLMs used as writing assistants before Swedish 2026 election — 1 src1.1 Study Audits Misalignment in Multi-Modal World Models — 1 src1.1 Retrieval-Grounded Reasoning Approach Proposed for Universal Multimodal Embeddings — 1 src4.0 Anthropic CEO Amodei calls for slower AI development and shared safety rules — 11 src2.2 Agility Robotics unveils Digit 5 humanoid for warehouses and factories — 2 src2.0 Apple ships rebuilt Siri with Google Gemini, but not in the EU — 2 src1.8 Siri AI in macOS 27 Golden Gate: FAQ, Germany availability, privacy questions — 2 src1.4 Sam Altman says OpenAI will not go public in 2026, citing AI safety concerns — 5 src1.1 OpenAI contractors review real ChatGPT conversations to rate responses, report says — 2 src1.1 Anthropic data retention policy prompts firms to limit Claude use for sensitive work — 1 src1.1 Study examines issue bias in LLMs used as writing assistants before Swedish 2026 election — 1 src1.1 Study Audits Misalignment in Multi-Modal World Models — 1 src1.1 Retrieval-Grounded Reasoning Approach Proposed for Universal Multimodal Embeddings — 1 src
HEATPULSEAI MAGAZINES
FLIP · FOLLOW · SAVE
Log inSign up

hyperparameter-tuning

topic6 events
papersTODAY 04:00 UTC

arXiv paper explores new directions for ACTS track-seeding parameter optimization

A new arXiv preprint examines ways to improve the ACTS software suite's handling of configuration parameters used in track seeding for charged-particle reconstruction. Because those settings strongly influence both reconstruction quality and computing cost, they are typically adjusted by hand through expert judgement and repeated testing. The work proposes new directions for reducing that manual tuning burden.

arxiv/particle-physics/optimization/open-source
papersTODAY 04:00 UTC

Study Analyzes 160,000 Training Runs to Improve Offline Policy Learning Baselines

A new arXiv paper examines how reporting choices, hyperparameter tuning, and dataset characteristics affect offline policy learning results. Drawing on roughly 160,000 training runs, the authors argue that reliable progress requires careful reporting, well-tuned baselines, and evaluation across varied conditions. The work offers practical guidance for making policy-learning benchmarks more reproducible and comparable.

reinforcement-learning/benchmarks/reproducibility/hyperparameters
papersSEP 12 04:00 UTC

ExpTest Uses Loss-Curve Hypothesis Testing to Pick Learning Rates Automatically

A research paper proposes ExpTest, a method that selects learning rates for deep neural networks on its own by statistically testing the shape of the training loss curve. The approach aims to reduce the manual searching and expensive grid searches that currently make hyperparameter tuning costly and less accessible. It is presented as part of ongoing work on arXiv in the cs.AI category.

research/training/hyperparameters/optimization
papersSEP 11 04:00 UTC

Paper Proposes Langevin Gradient Descent for Data-Driven Hyperparameter Tuning

A new arXiv preprint introduces the Langevin Gradient Descent Algorithm (LGD), which approximates the mean of a posterior distribution defined by a regression problem's loss function and regularizer in order to tune gradient descent hyperparameters. The authors frame this as a learning-to-learn problem and provide generalization guarantees for the approach. The work targets data-driven tuning of optimization settings rather than hand-selected values.

research/optimization/machine-learning/arxiv
papersSEP 11 04:00 UTC

arXiv paper targets lower-tail calibration of Gaussian processes for Bayesian optimization

An updated arXiv preprint proposes a goal-oriented approach to calibrating the lower tail of Gaussian process predictive distributions, which Bayesian optimization uses to choose where to evaluate costly objective functions. The abstract notes that kernel and hyperparameter choices strongly shape these predictions. The submission is a replacement version (v2) of an earlier preprint.

bayesian-optimization/gaussian-processes/calibration/arxiv
papersSEP 10 04:00 UTC

SCCM: Stream Cruise Control Method for Automated Drift Detection and Adaptation

A new arXiv paper introduces SCCM, a method for streaming machine learning that automatically detects concept drift, the shift in data distributions that degrades model performance over time. The approach seeks to keep predictive models accurate on evolving data while removing the dependency on fixed, manually tuned hyperparameters. The work was cross-listed between the cs.AI and cs.LG categories.

machine-learning/concept-drift/streaming/research