Results for “learning”

Trains and applies base-resolution deep learning models on ChIP-seq / ChIP-nexus / CUT&RUN data. Uses BPNet (Avsec 2021 Nat Genet 53:354; soft motif syntax from ChIP-nexus), chromBPNet (Pampari A et al 2025 Nat Genet; bias-factorized base-resolution profiles), EnFormer (Avsec 2021 Nat Methods 18:1196; 196 kb input, ~100 kb effective receptive field), DeepSEA (Zhou 2015; multi-task CNN), and JASPAR 2026 deep-learning collection (1259 BPNet ChIP models). Performs in silico mutagenesis for variant-effect prediction, DeepLIFT/Grad attribution, and TF-MoDISco motif discovery from attribution scores. Use when predicting variant effects on TF binding, discovering soft motif syntax / cooperativity, integrating ChIP-seq with sequence-only predictions, or applying precomputed JASPAR Deep Learning models to new variants.

The English summary is being prepared. Showing the original description.

Manage project learnings. Review, search, prune, and export what gstack has learned across sessions. Use when asked to "what have we learned", "show learnings", "prune stale learnings", or "export learnings". Proactively suggest when the user asks about past patterns or wonders "didn't we fix this before?"

The English summary is being prepared. Showing the original description.

Trains and applies base-resolution deep learning models on ChIP-seq / ChIP-nexus / CUT&RUN data. Uses BPNet (Avsec 2021 Nat Genet 53:354; soft motif syntax from ChIP-nexus), chromBPNet (Pampari A et al 2025 Nat Genet; bias-factorized base-resolution profiles), EnFormer (Avsec 2021 Nat Methods 18:1196; 196 kb input, ~100 kb effective receptive field), DeepSEA (Zhou 2015; multi-task CNN), and JASPAR 2026 deep-learning collection (1259 BPNet ChIP models). Performs in silico mutagenesis for variant-effect prediction, DeepLIFT/Grad attribution, and TF-MoDISco motif discovery from attribution scores. Use when predicting variant effects on TF binding, discovering soft motif syntax / cooperativity, integrating ChIP-seq with sequence-only predictions, or applying precomputed JASPAR Deep Learning models to new variants.

The English summary is being prepared. Showing the original description.

Add machine learning to an existing application through a structured workflow. The skill guides problem definition, data checks, separation from business logic, and training and evaluation of a simple initial model.

• Adding recommendations to an existing application
• Checking whether training data is ready
• Integrating a prediction model with fallback rules

Turn lessons from complex tasks, debugging, failures, or rollbacks into short learning entries. Capture the root cause, a reusable action, and a recognizable signal, then merge examples that share a cause with existing notes.

• Capture lessons after a complex task
• Record root causes and lessons from debugging
• Merge learning notes that share the same cause

Create agents that learn from interactions and adapt their behavior over time. Monitor learning progress, review efficiency and accuracy metrics, trigger adjustments from feedback, and share acquired knowledge with other agents through dedicated tools.

• Create an agent with initial learning settings
• Monitor learning progress, efficiency, and accuracy
• Adjust agent behavior based on feedback

Capture lessons from completed tasks, train patterns from explicit examples, and consolidate accumulated learning into long-term storage. Use SONA and MicroLoRA adaptation tools for individual domains, then verify results in the appropriate stores.

• Capturing lessons after a successful task
• Consolidating accumulated task histories
• Creating learning adapters for separate domains

Review recent Claude Code chats to identify development strengths, recurring difficulties, and learning priorities. Receive a personalized report with practical recommendations and relevant HackerNews articles, delivered to your own Slack direct messages.

• Reviewing development habits from recent chats
• Prioritizing recurring technical difficulties
• Finding learning articles relevant to current work

Learn AI engineering by building a project one stage at a time. The tutor guides predictions, learner-written code, automated checks, and reflection, then records progress so the next session can continue from the right stage.

• Learn AI engineering through hands-on projects
• Resume a project from the next unfinished stage
• Find coding mistakes with step-by-step hints

Build and evaluate single-agent reinforcement learning experiments, where an agent learns actions from rewards. Use Stable Baselines3 to choose algorithms, create environments, monitor training, and save models with the normalization data needed for consistent evaluation.

• Set up a reproducible reinforcement learning experiment
• Create and validate a custom Gymnasium environment
• Monitor training and save checkpoints

Plan and review PufferLib reinforcement learning workflows, where agents learn through rewards. Covers version-specific environment checks, parallel execution, training and evaluation plans, plus checkpoint inspection before loading saved models.

• Choose APIs for a specific PufferLib version
• Check environment behavior before training
• Create training and evaluation plans with separate seeds

Design subject lessons that also teach students how to learn. Add recall practice, explanations in their own words, predictions, and reflection to help learners check their understanding and plan later review while studying the main concepts.

• Adding recall practice to subject lessons
• Having students explain concepts in their own words
• Designing feedback and retries after mistakes

Prepare for an upcoming workday by combining Outlook meetings with open tasks and workspace context. Generate an HTML briefing with meeting preparation, scheduling conflicts, suggested priorities, and available time for learning and focused work.

• Prepare for tomorrow's meetings
• Spot overlapping meetings and scheduling issues
• Find time for learning and focused work

Build machine-learning trading signals from price and volume data. The workflow checks inputs, creates predictive features, and repeatedly trains on available historical data to produce bounded signals indicating predicted direction and strength.

• Building trading signals from price and volume data
• Comparing machine-learning models
• Testing expanding and sliding training windows

Build normal traffic profiles for industrial control networks and identify unusual connections, polling intervals, and protocol operations. Use machine learning and process records to investigate deviations and produce anomaly reports with severity levels.

• Establishing normal traffic profiles for industrial networks
• Detecting unusual connections and polling intervals
• Investigating Nozomi Guardian or Dragos alerts

Set up reinforcement learning, which trains models using rewards, with verl. Follow workflows for preparing data, configuring GRPO or PPO training, scaling across GPUs, and addressing memory errors or unstable training.

• Training a math reasoning model with GRPO
• Configuring PPO training with a critic model
• Scaling large model training across multiple nodes

Set up model training with torchforge for reinforcement learning, where rewards guide improvement. Covers training configurations, custom reward and loss functions, and distributed execution with Monarch and TorchTitan, keeping algorithm experiments separate from infrastructure concerns.

• Train mathematical reasoning models with GRPO
• Experiment with custom reward and loss functions
• Scale training across multiple GPUs or nodes

Guide language model training with TRL, from learning to follow instructions to matching human preferences. Covers method selection, training data, reward models that score responses, evaluation, and adjustments for memory limits or unstable training.

• Train a model on instruction and response examples
• Align responses using preferred and rejected answer pairs
• Build a training pipeline using human feedback

Explore training workflows that reduce harmful model responses using a written set of principles. The skill covers self-critique, revised training examples, and RLAIF, reinforcement learning from AI feedback, with implementation examples and hardware guidance.

• Designing principles for model safety training
• Creating training examples through self-critique and revision
• Planning reinforcement learning with AI-rated responses

Guides large language model reinforcement learning with miles, covering environment setup, low-precision training, consistency between training and inference, and speculative decoding, where a smaller model proposes tokens for a larger model to verify.

• Configure reinforcement learning for large MoE models
• Investigate instability in FP8 training
• Set up speculative decoding for faster rollout generation

Analyze single-cell data with scvi-tools, using deep learning to combine datasets, correct differences between experiments, and transfer cell labels. Model guides and reusable scripts cover RNA, protein, chromatin accessibility, spatial data, and downstream analysis.

• Integrate single-cell RNA datasets and correct batch effects
• Analyze paired RNA and protein or RNA and ATAC measurements
• Transfer cell-type labels from reference data

Choose Swedish videos and podcasts suited to your approximate language level and learning goals. Build a manageable plan for listening, reading, writing, and speaking, with specific resources and short exercises to practice each skill.

• Choosing Swedish resources suited to your level
• Planning podcast practice for a daily commute
• Practicing pronunciation by repeating video dialogue

Plan smart contract audits using research from security organizations. Combine threat modeling, static code analysis, fuzzing that varies test inputs, and manual review, while using attack patterns and a four-phase roadmap to guide further learning.

• Planning a smart contract security audit
• Setting up Echidna or Medusa fuzzing
• Reviewing attack patterns and past audit findings

Guides large language model reinforcement learning with miles, including MoE models that route work to specialized components. Covers low-precision training, matching training and inference behavior, and accelerating training response generation with draft models.

• Setting up reinforcement learning for large MoE models
• Configuring FP8 or INT4 training
• Checking consistency between training and inference