Mike A. Merrill

Mike A. Merrill, Tim Althoff · uw

Detailed mobile sensing data from phones, watches, and fitness trackers offer an unparalleled opportunity to quantify and act upon previously unmeasurable behavioral changes in order to improve individual health and accelerate responses to emerging diseases. Unlike in natural language processing and computer vision, deep representation learning has yet to broadly impact this domain, in which the vast majority of research and clinical applications still rely on manually defined features and boosted tree models or even forgo predictive modeling altogether due to insufficient accuracy. This is due to unique challenges in the behavioral health domain, including very small datasets (~10^1 participants), which frequently contain missing data, consist of long time series with critical long-range dependencies (length>10^4), and extreme class imbalances (>10^3:1). Here, we introduce a neural architecture for multivariate time series classification designed to address these unique domain challenges. Our proposed behavioral representation learning approach combines novel tasks for self-supervised pretraining and transfer learning to address data scarcity, and captures long-range dependencies across long-history time series through transformer self-attention following convolutional neural network-based dimensionality reduction. We propose an evaluation framework aimed at reflecting expected real-world performance in plausible deployment scenarios. Concretely, we demonstrate (1) performance improvements over baselines of up to 0.15 ROC AUC across five prediction tasks, (2) transfer learning-induced performance improvements of 16% PR AUC in small data scenarios, and (3) the potential of transfer learning in novel disease scenarios through an exploratory case study of zero-shot COVID-19 prediction in an independent data set. Finally, we discuss potential implications for medical surveillance testing.

Ken Gu, Ruoxi Shang, Ruien Jiang et al. · uw

Data-driven scientific discovery requires the iterative integration of scientific domain knowledge, statistical expertise, and an understanding of data semantics to make nuanced analytical decisions, e.g., about which variables, transformations, and statistical models to consider. LM-based agents equipped with planning, memory, and code execution capabilities have the potential to support data-driven science. However, evaluating agents on such open-ended tasks is challenging due to multiple valid approaches, partially correct steps, and different ways to express the same decisions. To address these challenges, we present BLADE, a benchmark to automatically evaluate agents' multifaceted approaches to open-ended research questions. BLADE consists of 12 datasets and research questions drawn from existing scientific literature, with ground truth collected from independent analyses by expert data scientists and researchers. To automatically evaluate agent responses, we developed corresponding computational methods to match different representations of analyses to this ground truth. Though language models possess considerable world knowledge, our evaluation shows that they are often limited to basic analyses. However, agents capable of interacting with the underlying data demonstrate improved, but still non-optimal, diversity in their analytical decision making. Our work enables the evaluation of agents for data-driven science and provides researchers deeper insights into agents' analysis approaches.

Mike A. Merrill, Mingtian Tan, Vinayak Gupta et al.

Time series are critical for decision-making in fields like finance and healthcare. Their importance has driven a recent influx of works passing time series into language models, leading to non-trivial forecasting on some datasets. But it remains unknown whether non-trivial forecasting implies that language models can reason about time series. To address this gap, we generate a first-of-its-kind evaluation framework for time series reasoning, including formal tasks and a corresponding dataset of multi-scale time series paired with text captions across ten domains. Using these data, we probe whether language models achieve three forms of reasoning: (1) Etiological Reasoning - given an input time series, can the language model identify the scenario that most likely created it? (2) Question Answering - can a language model answer factual questions about time series? (3) Context-Aided Forecasting - does highly relevant textual context improve a language model's time series forecasts? We find that otherwise highly-capable language models demonstrate surprisingly limited time series reasoning: they score marginally above random on etiological and question answering tasks (up to 30 percentage points worse than humans) and show modest success in using context to improve forecasting. These weakness showcase that time series reasoning is an impactful, yet deeply underdeveloped direction for language model research. We also make our datasets and code public at to support further research in this direction at https://github.com/behavioral-data/TSandLanguage

Etash Guha, Ryan Marten, Sedrick Keh et al. · cmu

Reasoning models have made rapid progress on many benchmarks involving math, code, and science. Yet, there are still many open questions about the best training recipes for reasoning since state-of-the-art models often rely on proprietary datasets with little to no public information available. To address this, the goal of the OpenThoughts project is to create open-source datasets for training reasoning models. After initial explorations, our OpenThoughts2-1M dataset led to OpenThinker2-32B, the first model trained on public reasoning data to match DeepSeek-R1-Distill-32B on standard reasoning benchmarks such as AIME and LiveCodeBench. We then improve our dataset further by systematically investigating each step of our data generation pipeline with 1,000+ controlled experiments, which led to OpenThoughts3. Scaling the pipeline to 1.2M examples and using QwQ-32B as teacher yields our OpenThoughts3-7B model, which achieves state-of-the-art results: 53% on AIME 2025, 51% on LiveCodeBench 06/24-01/25, and 54% on GPQA Diamond - improvements of 15.3, 17.2, and 20.5 percentage points compared to the DeepSeek-R1-Distill-Qwen-7B. All of our datasets and models are available on https://openthoughts.ai.

Mingtian Tan, Mike A. Merrill, Zack Gottesman et al.

Time series data measure how environments change over time and drive decision-making in critical domains like finance and healthcare. A common goal in analyzing time series data is to understand the underlying events that cause the observed variations. We conduct the first study of whether Large Language Models (LLMs) can infer events described with natural language from time series data. We evaluate 18 LLMs on a task to match event sequences with real-valued time series data using a new benchmark we develop using sports data. Several current LLMs demonstrate promising abilities, with OpenAI's o1 performing the best but with DS-R1-distill-Qwen-32B outperforming proprietary models such as GPT-4o. From insights derived from analyzing reasoning failures, we also find clear avenues to improve performance. By applying post-training optimizations, i.e., distillation and self-improvement, we significantly enhance the performance of the Qwen2.5 1.5B, achieving results second only to o1. All resources needed to reproduce our work are available: https://github.com/BennyTMT/GAMETime

Mingtian Tan, Mike A. Merrill, Vinayak Gupta et al.

Large language models (LLMs) are being applied to time series forecasting. But are language models actually useful for time series? In a series of ablation studies on three recent and popular LLM-based time series forecasting methods, we find that removing the LLM component or replacing it with a basic attention layer does not degrade forecasting performance -- in most cases, the results even improve! We also find that despite their significant computational cost, pretrained LLMs do no better than models trained from scratch, do not represent the sequential dependencies in time series, and do not assist in few-shot settings. Additionally, we explore time series encoders and find that patching and attention structures perform similarly to LLM-based forecasters.

Mike A. Merrill, Akshay Paruchuri, Naghmeh Rezaei et al.

Deriving personalized insights from popular wearable trackers requires complex numerical reasoning that challenges standard LLMs, necessitating tool-based approaches like code generation. Large language model (LLM) agents present a promising yet largely untapped solution for this analysis at scale. We introduce the Personal Health Insights Agent (PHIA), a system leveraging multistep reasoning with code generation and information retrieval to analyze and interpret behavioral health data. To test its capabilities, we create and share two benchmark datasets with over 4000 health insights questions. A 650-hour human expert evaluation shows that PHIA significantly outperforms a strong code generation baseline, achieving 84% accuracy on objective, numerical questions and, for open-ended ones, earning 83% favorable ratings while being twice as likely to achieve the highest quality rating. This work can advance behavioral health by empowering individuals to understand their data, enabling a new era of accessible, personalized, and data-driven wellness for the wider population.

Mike A. Merrill, Tim Althoff

While deep learning has revolutionized research and applications in NLP and computer vision, this has not yet been the case for behavioral modeling and behavioral health applications. This is because the domain's datasets are smaller, have heterogeneous datatypes, and typically exhibit a large degree of missingness. Therefore, off-the-shelf deep learning models require significant, often prohibitive, adaptation. Accordingly, many research applications still rely on manually coded features with boosted tree models, sometimes with task-specific features handcrafted by experts. Here, we address these challenges by providing a neural architecture framework for mobile sensing data that can learn generalizable feature representations from time series and demonstrates the feasibility of transfer learning on small data domains through finetuning. This architecture combines benefits from CNN and Trans-former architectures to (1) enable better prediction performance by learning directly from raw minute-level sensor data without the need for handcrafted features by up to 0.33 ROC AUC, and (2) use pretraining to outperform simpler neural models and boosted decision trees with data from as few a dozen participants.

Ge Zhang, Mike A. Merrill, Yang Liu et al.

Large scale analysis of source code, and in particular scientific source code, holds the promise of better understanding the data science process, identifying analytical best practices, and providing insights to the builders of scientific toolkits. However, large corpora have remained unanalyzed in depth, as descriptive labels are absent and require expert domain knowledge to generate. We propose a novel weakly supervised transformer-based architecture for computing joint representations of code from both abstract syntax trees and surrounding natural language comments. We then evaluate the model on a new classification task for labeling computational notebook cells as stages in the data analysis process from data import to wrangling, exploration, modeling, and evaluation. We show that our model, leveraging only easily-available weak supervision, achieves a 38% increase in accuracy over expert-supplied heuristics and outperforms a suite of baselines. Our model enables us to examine a set of 118,000 Jupyter Notebooks to uncover common data analysis patterns. Focusing on notebooks with relationships to academic articles, we conduct the largest ever study of scientific code and find that notebook composition correlates with the citation count of corresponding papers.