Data & Analytics

Research prediction markets as data sources or oracle signals for products, agents, dashboards, and corporate decision intelligence. Use for source-grounded analysis of market-implied probabilities, caveats, and integration patterns without investment advice.

Builds an XLSX competitor-landscape workbook and a brand-matched PDF report from scratch: research sweep, category taxonomy, feature-gap analysis, computed market-pattern insights, and sales battlecards. Invoke on phrases like build a competitor comparison spreadsheet, research our competitors, make a battlecard deck, competitive landscape report, or feature gap analysis against competitors.

> **Source**: `../role-data-scientist/SKILL.md` > > This file is a reference. The main skill definition is in the `.claude/skills/` directory.

Discover that generative models and vision encoders share an underlying approximately Gaussian latent space (UNE). Enable controllable image editing by operating on linear directions in diffusion noise space, revealing that semantic information is present without explicit supervision.

Introduce Quasicryth, a text compressor using Fibonacci quasicrystal tilings for phrase-level compression. Prove that aperiodic structures never structurally collapse at depth, enabling compression at arbitrary hierarchy levels. Bridge quasicrystal mathematics with practical compression, achieving 22.59% ratio on enwik9 with unbounded scaling advantages.

Quantify confidence in video generation by estimating latent-space uncertainty and mapping to interpretable heatmaps. Detect untrustworthy regions in generated frames for robot planning and policy evaluation—critical when video hallucinations could cause real-world failures.

Improve long-horizon world model fidelity using RL with clip-level rollouts and complementary reward functions for action accuracy and visual quality. Breaks computational constraints by evaluating candidate clips incrementally rather than full sequences, enabling efficient multi-objective optimization.

Break through supervised fine-tuning saturation by mixing logits from weaker model checkpoints into strong model training targets. Amplifies hard negatives that strong models have already suppressed, enabling continued learning after standard training plateaus.

VCRL improves LLM mathematical reasoning by dynamically adjusting training sample difficulty based on group reward variance. Uses variance-based sampling with memory replay to focus on moderately-difficult samples where models succeed ~50% of the time, achieving 4.67-point improvement over GRPO on 8B models.

Route reasoning between discrete token space (when uncertain) and latent soft embeddings (when confident). Use maximum next-token probability as a routing threshold to dynamically select the reasoning space, improving accuracy under latent reasoning while reducing computational cost through selective discrete sampling.

Generate coherent, consistent long-form text through structured planning, hierarchical reflection, and Monte Carlo tree search-guided optimization.

Train process reward models without ground-truth references using synthetic verification data from generators and verifiers. SPARK achieves 67.5 F1 on ProcessBench—ideal when step-level annotations are expensive but verification is available.

Enable MLLMs to handle visually degraded images by explicitly reasoning about degradation types and severity. Trains models to perceive degradation parameters, analyze semantic impact, and reconstruct interpretations while adapting reasoning depth to degradation complexity—achieving robust understanding with interpretable explanations.

Dual-scoring mechanism (Magnitude and Stability scores) enabling process-level rewards that penalize inefficient reasoning patterns like overthinking and backtracking without trained reward models. Improve reasoning efficiency in RL without additional supervision.

Improve LLM training stability and generalization by reparameterizing weight matrices as orthogonal transformations, achieving better perplexity than AdamW with fewer trainable parameters.

Accelerate diffusion transformers through training-free sparse attention combining exact computation for critical blocks with efficient approximation for non-critical ones. Achieves 2-3× speedup without retraining.

Scale distributed LLM training without all-reduce synchronization using dynamic pipeline routing and modified Nesterov momentum, achieving 4% faster convergence than DiLoCo with exponentially lower communication.

Create 7.5M long-form mathematical solution traces with multi-mode supervision (high/medium/low reasoning depths, with/without Python). Integrate 85K competition problems (AoPS) and 262K community questions (StackExchange). Implement sequential bucketing training achieving 2-3× speedup while maintaining accuracy.

Apply μ-parametrization to Mixture-of-Experts architectures to enable reliable hyperparameter transfer across model sizes, eliminating costly retuning when scaling to trillion-parameter systems.

Determine optimal MoE sparsity by separating memorization and reasoning trade-offs: active FLOPs improve reasoning while total parameters improve memorization, requiring joint optimization

Trains mean-velocity models on rectified couplings from pretrained flow models to dramatically smooth loss landscape, enabling faster convergence and superior one-step generation quality without additional training data.

Adapt DPO temperature parameters per sample using reward model margins, amplifying learning signals for hard preference pairs while dampening easy ones.

Preserve reasoning diversity in RL by protecting low-probability tokens that represent novel thinking paths. Trigger: maintain exploration capability during RL training while avoiding entropy collapse on reasoning tasks.

Train reward models to evaluate long-context responses by introducing faithfulness as a criterion and using consistency-based alignment to maintain judgment-explanation coherence.