Blog

Introduction to Double/Debiased Machine Learning

Causal Inference
Machine Learning
Econometrics

DML combines Neyman orthogonal scores with cross-fitting to enable valid statistical inference about low-dimensional causal parameters while using flexible ML methods to estimate high-dimensional nuisance functions. This review covers the theory, two simulation illustrations, and three empirical applications including staggered adoption designs and non-tabular text data.

Mar 7, 2026
14 min

Kernel Debiased Estimation: Semiparametric Efficiency Without Deriving the Influence Function

Causal Inference
Semiparametric Statistics
Kernel Methods

ULFS-KDPE constructs a single data-adaptive distributional flow in a reproducing kernel Hilbert space that simultaneously debiases all pathwise differentiable parameters, achieving semiparametric efficiency without ever requiring an explicit influence function.

Mar 6, 2026
16 min

Perturbed Double Machine Learning: Valid Inference When Your Nuisance Estimators Are Too Slow

Causal Inference
Machine Learning
Semiparametric Statistics

Standard double machine learning requires nuisance estimators that converge at n^{-1/4} or faster for valid confidence intervals. Zheng, Bonvini, and Guo show how to get honest coverage even when that rate is unattainable by injecting calibrated noise, re-estimating many times, and keeping only the perturbations that behave well.

Mar 5, 2026
15 min

Multi-Agent AI Patterns: A Developer’s Field Guide

AI Agents
Design Patterns
LLMs
Software Architecture

Suman Das catalogs the eight core multi-agent design patterns every AI developer needs — from sequential pipelines to evaluator-optimizer loops — with clear guidance on when to use each, which frameworks support them, and how production systems combine them.

Mar 4, 2026
27 min

Agentic AI for Alzheimer’s: Automating Hypothesis Testing at Scale

AI Agents
Healthcare
Alzheimer's
Biomedical Research

Saichandran et al. from Boston University build a multi-agent system with four specialized LLM agents that automatically tests Alzheimer’s disease hypotheses against the NACC cohort (n=48,876). The system verified 42 of 100 ADRD hypotheses while correctly rejecting 85% of negative controls.

Mar 3, 2026
10 min

Star-Agents: Using LLM Agents to Optimize Training Data

LLMs
Data Optimization
Instruction Tuning
NeurIPS

Zhou et al.’s Star-Agents system uses multiple LLM agent pairs, dual-model evaluation, and evolutionary sampling to automatically generate and curate high-quality instruction-tuning data — achieving 12% average improvement and up to 40% on Fermi tasks.

Mar 2, 2026
12 min

From Features to Actions: Rethinking Explainability for AI Agents

AI Agents
Explainability
XAI
LLMs

Chaduvula et al. from the Vector Institute and Mayo Clinic argue that traditional XAI tools (SHAP, LIME) break down for agentic AI systems. Their Minimal Explanation Packet framework and TAU-bench evaluation reveal that state tracking consistency — not individual feature importance — is the strongest predictor of agent failure.

Mar 1, 2026
14 min

HypoGeniC: Teaching LLMs to Generate Hypotheses from Data

LLMs
Hypothesis Generation
Machine Learning
Data Science

Zhou et al.’s HypoGeniC uses a UCB-inspired algorithm to iteratively generate hypotheses from data using LLMs — outperforming zero-shot by up to 60% and matching fine-tuned RoBERTa on some tasks, all without gradient updates.

Feb 28, 2026
11 min

Non-Manipulable Causes: Pearl’s Answer to Causal Inference’s Oldest Objection

Causal Inference
Statistics
Philosophy of Science

Judea Pearl’s 2018 paper tackles the ‘no causation without manipulation’ objection head-on — showing that structural causal models can define causal effects for inherently non-manipulable variables like race, gender, or obesity through the do-operator, without requiring physical intervention.

Feb 27, 2026
8 min

Computer Age Statistical Inference: The Grand Tour of Modern Statistics

Statistics
Machine Learning
Inference
Book Review

Efron and Hastie’s masterwork traces 250 years of statistical inference into the computer age — from Bayes and Fisher through bootstrapping, LASSO, and deep learning. A book review of the text that bridges classical theory and modern data science.

Feb 26, 2026
9 min

ReAct: When Language Models Learn to Think and Do

AI Agents
Reasoning
LLMs
ICLR

Yao et al.’s ICLR 2023 paper introduced ReAct — interleaving reasoning traces with actions in language models. The key insight: chain-of-thought alone hallucinates (56% error rate), but grounding reasoning in real actions nearly eliminates it. ReAct became the blueprint for modern AI agents.

Feb 25, 2026
8 min

MLAgentBench: Can AI Agents Actually Do Machine Learning Research?

AI Agents
Benchmarking
Machine Learning
ICLR

Huang et al.’s ICLR 2024 benchmark puts LLM agents to the test on real ML research tasks — from CIFAR-10 to Kaggle competitions. GPT-4 handles canonical tasks (~90%) but struggles badly with open-ended research (0-30% on Kaggle).

Feb 24, 2026
9 min

Plan-and-Solve Prompting: Fixing the Gaps in Chain-of-Thought

LLMs
Prompting
Reasoning
ACL

Wang et al. identified three systematic failure modes in chain-of-thought prompting — missing steps, semantic misunderstanding, and calculation errors — then designed PS+ prompting to address them. The result: consistent gains over Zero-shot-CoT across arithmetic, commonsense, and symbolic reasoning.

Feb 23, 2026
8 min

Chain-of-Thought Prompting: Teaching LLMs to Show Their Work

LLMs
Prompting
Reasoning
NeurIPS

Wei et al.’s landmark NeurIPS 2022 paper showed that simply asking large language models to reason step-by-step unlocks dramatic improvements in arithmetic, commonsense, and symbolic reasoning — but only above ~100B parameters.

Feb 22, 2026
11 min

Evaluating AGENTS.md: Do Repository-Level Context Files Actually Help Coding Agents?

AI Agents
Software Engineering
LLMs
Benchmarking

Anthropic’s empirical evaluation of AGENTS.md files — structured instructions that tell coding agents how to navigate a repository. The results: a 7.0% improvement on SWE-bench, but the story is more nuanced than the headline.

Feb 21, 2026
11 min

Estimation and Inference of Heterogeneous Treatment Effects Using Random Forests

Causal Inference
Machine Learning
Random Forests
Treatment Effects

Breaking down Wager & Athey’s causal forest — how they turned random forests into a principled tool for estimating who benefits most from a treatment, complete with valid confidence intervals.

Feb 20, 2026
10 min

AI Job Loss and the Limits of Comparative Advantage

AI
Economics
Labor Markets
Comparative Advantage

Unpacking claywren’s argument on LessWrong about why the classical economic case for human employment may not survive superhuman AI — and what the comparative advantage framework gets wrong.

Feb 19, 2026
11 min

Estimating Average Treatment Effects Under Unconfoundedness

Causal Inference
Econometrics
Program Evaluation
Propensity Score

Breaking down Imbens & Wooldridge’s influential lecture notes on matching, weighting, and subclassification methods for estimating treatment effects when selection is on observables.

Feb 18, 2026
11 min

McKenzie Told You to Add More Survey Rounds — But His Power Formula Was Wrong

Experimental Design
Causal Inference
Panel Data
Paper Review

Burlig, Preonas & Woerman show that the constant-autocorrelation assumption behind standard panel power calculations leads to overpowered short panels and underpowered long panels. Their fix: Serial-Correlation-Robust power formulas.

Feb 17, 2026
15 min

Your RCT Has a Power Problem — And the Fix Is More Surveys, Not More People

Experimental Design
Causal Inference
Development Economics
Paper Review

A deep dive into McKenzie’s classic paper on why most experiments waste statistical power by collecting only one baseline and one follow-up. For noisy outcomes like profits and income, adding measurement rounds beats adding sample size.

Feb 16, 2026
14 min

You Have the Whole Population — Why Are You Still Computing Standard Errors?

Causal Inference
Econometrics
Statistics
Paper Review

A deep dive into the Abadie-Athey-Imbens-Wooldridge paper that asks a question most data scientists never think to ask: where does your uncertainty actually come from? The answer changes everything about how you report standard errors.

Feb 15, 2026
18 min

Your Pricing Model Is Learning the Manager, Not the Market — How to Fix Multi-Task Demand Estimation

Causal Inference
Pricing
Meta-Learning
Econometrics
Paper Review

A deep dive into DCMOML — a new framework that fixes the hidden confounding problem in multi-task demand learning. When historical prices reflect human decisions, standard meta-learners learn the policy, not the demand curve.

Feb 14, 2026
24 min

No Tuning, No Task Cues, No Problem — IBM’s Dead-Simple Approach to Automatic Prompt Engineering

Prompt Engineering
LLMs
Automation
Paper Review

A deep dive into a paper from IBM Research that generates effective prompts with just 8 examples, zero tuning, and zero extra LLM calls for scoring — and still matches DSPy on benchmarks.

Feb 13, 2026
17 min

58 Ways to Talk to an LLM — A Field Guide to Prompt Engineering

Prompt Engineering
LLMs
NLP
Paper Review

A deep dive into ‘The Prompt Report’ — the most comprehensive taxonomy of prompt engineering techniques to date. 58 text-based techniques, 40 multimodal methods, and a reality check on what actually works.

Feb 12, 2026
13 min

Your LLM Is Smart — But Can It Price a Deal? The Case for ML-as-a-Tool

Machine Learning
LLMs
AI Architecture
Paper Review

A deep dive into the MLAT framework — where traditional ML models become callable tools inside LLM agent workflows. Featuring PitchCraft, a system that turned 3-hour sales proposals into 10-minute automated pipelines.

Feb 11, 2026
11 min
No matching items