Active Task Disambiguation with LLMs

Despite the impressive performance of large language models (LLMs) across various benchmarks, their ability to address ambiguously specified problems--frequent in real-world interactions--remains underexplored. To address this gap, we introduce a formal definition of task ambiguity and frame the problem of task disambiguation through the lens of Bayesian Experimental Design. By posing clarifying questions, LLM agents can acquire additional task specifications, progressively narrowing the space of viable solutions and reducing the risk of generating unsatisfactory outputs. Yet, generating effective clarifying questions requires LLM agents to engage in a form of meta-cognitive reasoning, an ability LLMs may presently lack. Our proposed approach of active task disambiguation enables LLM agents to generate targeted questions maximizing the information gain. Effectively, this approach shifts the load from implicit to explicit reasoning about the space of viable solutions. Empirical results demonstrate that this form of question selection leads to more effective task disambiguation in comparison to approaches relying on reasoning solely within the space of questions.

Reusing Embeddings: Reproducible Reward Model Research in Large Language Model Alignment without GPUs

Large Language Models (LLMs) have made substantial strides in structured tasks through Reinforcement Learning (RL), demonstrating proficiency in mathematical reasoning and code generation. However, applying RL in broader domains like chatbots and content generation -- through the process known as Reinforcement Learning from Human Feedback (RLHF) -- presents unique challenges. Reward models in RLHF are critical, acting as proxies that evaluate the alignment of LLM outputs with human intent. Despite advancements, the development of reward models is hindered by challenges such as computational heavy training, costly evaluation, and therefore poor reproducibility. We advocate for using embedding-based input in reward model research as an accelerated solution to those challenges. By leveraging embeddings for reward modeling, we can enhance reproducibility, reduce computational demands on hardware, improve training stability, and significantly reduce training and evaluation costs, hence facilitating fair and efficient comparisons in this active research area. We then show a case study of reproducing existing reward model ensemble research using embedding-based reward models. We discussed future avenues for research, aiming to contribute to safer and more effective LLM deployments.

No Equations Needed: Learning System Dynamics Without Relying on Closed-Form ODEs

Data-driven modeling of dynamical systems is a crucial area of machine learning. In many scenarios, a thorough understanding of the model's behavior becomes essential for practical applications. For instance, understanding the behavior of a pharmacokinetic model, constructed as part of drug development, may allow us to both verify its biological plausibility (e.g., the drug concentration curve is non-negative and decays to zero) and to design dosing guidelines. Discovery of closed-form ordinary differential equations (ODEs) can be employed to obtain such insights by finding a compact mathematical equation and then analyzing it (a two-step approach). However, its widespread use is currently hindered because the analysis process may be time-consuming, requiring substantial mathematical expertise, or even impossible if the equation is too complex. Moreover, if the found equation's behavior does not satisfy the requirements, editing it or influencing the discovery algorithms to rectify it is challenging as the link between the symbolic form of an ODE and its behavior can be elusive. This paper proposes a conceptual shift to modeling low-dimensional dynamical systems by departing from the traditional two-step modeling process. Instead of first discovering a closed-form equation and then analyzing it, our approach, direct semantic modeling, predicts the semantic representation of the dynamical system (i.e., description of its behavior) directly from data, bypassing the need for complex post-hoc analysis. This direct approach also allows the incorporation of intuitive inductive biases into the optimization algorithm and editing the model's behavior directly, ensuring that the model meets the desired specifications. Our approach not only simplifies the modeling pipeline but also enhances the transparency and flexibility of the resulting models compared to traditional closed-form ODEs.

Towards Human-Guided, Data-Centric LLM Co-Pilots

Machine learning (ML) has the potential to revolutionize healthcare, but its adoption is often hindered by the disconnect between the needs of domain experts and translating these needs into robust and valid ML tools. Despite recent advances in LLM-based co-pilots to democratize ML for non-technical domain experts, these systems remain predominantly focused on model-centric aspects while overlooking critical data-centric challenges. This limitation is problematic in complex real-world settings where raw data often contains complex issues, such as missing values, label noise, and domain-specific nuances requiring tailored handling. To address this we introduce CliMB-DC, a human-guided, data-centric framework for LLM co-pilots that combines advanced data-centric tools with LLM-driven reasoning to enable robust, context-aware data processing. At its core, CliMB-DC introduces a novel, multi-agent reasoning system that combines a strategic coordinator for dynamic planning and adaptation with a specialized worker agent for precise execution. Domain expertise is then systematically incorporated to guide the reasoning process using a human-in-the-loop approach. To guide development, we formalize a taxonomy of key data-centric challenges that co-pilots must address. Thereafter, to address the dimensions of the taxonomy, we integrate state-of-the-art data-centric tools into an extensible, open-source architecture, facilitating the addition of new tools from the research community. Empirically, using real-world healthcare datasets we demonstrate CliMB-DC's ability to transform uncurated datasets into ML-ready formats, significantly outperforming existing co-pilot baselines for handling data-centric challenges. CliMB-DC promises to empower domain experts from diverse domains -- healthcare, finance, social sciences and more -- to actively participate in driving real-world impact using ML.

Few-shot Steerable Alignment: Adapting Rewards and LLM Policies with Neural Processes

As large language models (LLMs) become increasingly embedded in everyday applications, ensuring their alignment with the diverse preferences of individual users has become a critical challenge. Currently deployed approaches typically assume homogeneous user objectives and rely on single-objective fine-tuning. However, human preferences are inherently heterogeneous, influenced by various unobservable factors, leading to conflicting signals in preference data. Existing solutions addressing this diversity often require costly datasets labelled for specific objectives and involve training multiple reward models or LLM policies, which is computationally expensive and impractical. In this work, we present a novel framework for few-shot steerable alignment, where users' underlying preferences are inferred from a small sample of their choices. To achieve this, we extend the Bradley-Terry-Luce model to handle heterogeneous preferences with unobserved variability factors and propose its practical implementation for reward modelling and LLM fine-tuning. Thanks to our proposed approach of functional parameter-space conditioning, LLMs trained with our framework can be adapted to individual preferences at inference time, generating outputs over a continuum of behavioural modes. We empirically validate the effectiveness of methods, demonstrating their ability to capture and align with diverse human preferences in a data-efficient manner. Our code is made available at: https://github.com/kasia-kobalczyk/few-shot-steerable-alignment.

LLMs for Generalizable Language-Conditioned Policy Learning under Minimal Data Requirements

To develop autonomous agents capable of executing complex, multi-step decision-making tasks as specified by humans in natural language, existing reinforcement learning approaches typically require expensive labeled datasets or access to real-time experimentation. Moreover, conventional methods often face difficulties in generalizing to unseen goals and states, thereby limiting their practical applicability. This paper presents TEDUO, a novel training pipeline for offline language-conditioned policy learning. TEDUO operates on easy-to-obtain, unlabeled datasets and is suited for the so-called in-the-wild evaluation, wherein the agent encounters previously unseen goals and states. To address the challenges posed by such data and evaluation settings, our method leverages the prior knowledge and instruction-following capabilities of large language models (LLMs) to enhance the fidelity of pre-collected offline data and enable flexible generalization to new goals and states. Empirical results demonstrate that the dual role of LLMs in our framework-as data enhancers and generalizers-facilitates both effective and data-efficient learning of generalizable language-conditioned policies.

Quantifying perturbation impacts for large language models

We consider the problem of quantifying how an input perturbation impacts the outputs of large language models (LLMs), a fundamental task for model reliability and post-hoc interpretability. A key obstacle in this domain is disentangling the meaningful changes in model responses from the intrinsic stochasticity of LLM outputs. To overcome this, we introduce Distribution-Based Perturbation Analysis (DBPA), a framework that reformulates LLM perturbation analysis as a frequentist hypothesis testing problem. DBPA constructs empirical null and alternative output distributions within a low-dimensional semantic similarity space via Monte Carlo sampling. Comparisons of Monte Carlo estimates in the reduced dimensionality space enables tractable frequentist inference without relying on restrictive distributional assumptions. The framework is model-agnostic, supports the evaluation of arbitrary input perturbations on any black-box LLM, yields interpretable p-values, supports multiple perturbation testing via controlled error rates, and provides scalar effect sizes for any chosen similarity or distance metric. We demonstrate the effectiveness of DBPA in evaluating perturbation impacts, showing its versatility for perturbation analysis.

Unlocking Historical Clinical Trial Data with ALIGN: A Compositional Large Language Model System for Medical Coding

The reuse of historical clinical trial data has significant potential to accelerate medical research and drug development. However, interoperability challenges, particularly with missing medical codes, hinders effective data integration across studies. While Large Language Models (LLMs) offer a promising solution for automated coding without labeled data, current approaches face challenges on complex coding tasks. We introduce ALIGN, a novel compositional LLM-based system for automated, zero-shot medical coding. ALIGN follows a three-step process: (1) diverse candidate code generation; (2) self-evaluation of codes and (3) confidence scoring and uncertainty estimation enabling human deferral to ensure reliability. We evaluate ALIGN on harmonizing medication terms into Anatomical Therapeutic Chemical (ATC) and medical history terms into Medical Dictionary for Regulatory Activities (MedDRA) codes extracted from 22 immunology trials. ALIGN outperformed the LLM baselines, while also providing capabilities for trustworthy deployment. For MedDRA coding, ALIGN achieved high accuracy across all levels, matching RAG and excelling at the most specific levels (87-90% for HLGT). For ATC coding, ALIGN demonstrated superior performance, particularly at lower hierarchy levels (ATC Level 4), with 72-73% overall accuracy and 86-89% accuracy for common medications, outperforming baselines by 7-22%. ALIGN's uncertainty-based deferral improved accuracy by 17% to 90% accuracy with 30% deferral, notably enhancing performance on uncommon medications. ALIGN achieves this cost-efficiently at \$0.0007 and \$0.02 per code for GPT-4o-mini and GPT-4o, reducing barriers to clinical adoption. ALIGN advances automated medical coding for clinical trial data, contributing to enhanced data interoperability and reusability, positioning it as a promising tool to improve clinical research and accelerate drug development.

Quantifying Aleatoric Uncertainty of the Treatment Effect: A Novel Orthogonal Learner

Estimating causal quantities from observational data is crucial for understanding the safety and effectiveness of medical treatments. However, to make reliable inferences, medical practitioners require not only estimating averaged causal quantities, such as the conditional average treatment effect, but also understanding the randomness of the treatment effect as a random variable. This randomness is referred to as aleatoric uncertainty and is necessary for understanding the probability of benefit from treatment or quantiles of the treatment effect. Yet, the aleatoric uncertainty of the treatment effect has received surprisingly little attention in the causal machine learning community. To fill this gap, we aim to quantify the aleatoric uncertainty of the treatment effect at the covariate-conditional level, namely, the conditional distribution of the treatment effect (CDTE). Unlike average causal quantities, the CDTE is not point identifiable without strong additional assumptions. As a remedy, we employ partial identification to obtain sharp bounds on the CDTE and thereby quantify the aleatoric uncertainty of the treatment effect. We then develop a novel, orthogonal learner for the bounds on the CDTE, which we call AU-learner. We further show that our AU-learner has several strengths in that it satisfies Neyman-orthogonality and is doubly robust. Finally, we propose a fully-parametric deep learning instantiation of our AU-learner.

Autoformulation of Mathematical Optimization Models Using LLMs

Mathematical optimization is fundamental to decision-making across diverse domains, from operations research to healthcare. Yet, translating real-world problems into optimization models remains a formidable challenge, often demanding specialized expertise. This paper formally introduces the concept of $\textbf{autoformulation}$ -- an automated approach to creating optimization models from natural language descriptions for commercial solvers. We identify the three core challenges of autoformulation: (1) defining the vast, problem-dependent hypothesis space, (2) efficiently searching this space under uncertainty, and (3) evaluating formulation correctness (ensuring a formulation accurately represents the problem). To address these challenges, we introduce a novel method leveraging $\textit{Large Language Models}$ (LLMs) within a $\textit{Monte-Carlo Tree Search}$ framework. This approach systematically explores the space of possible formulations by exploiting the hierarchical nature of optimization modeling. LLMs serve two key roles: as dynamic formulation hypothesis generators and as evaluators of formulation correctness. To enhance search efficiency, we introduce a pruning technique to remove trivially equivalent formulations. Empirical evaluations across benchmarks containing linear and mixed-integer programming problems demonstrate our method's superior performance. Additionally, we observe significant efficiency gains from employing LLMs for correctness evaluation and from our pruning techniques.