AI conference papers

AI conference paper library

Search papers from 30+ top conferences, including NeurIPS, ICLR, ICML, CVPR, ACL and AAAI. Filter by keyword, then open any paper for its abstract, authors and links.

arXiv

Optimality of Sub-network Laplace Approximations: New Results and Methods

Swarnali Raha, Kshitij Khare, Rohit K Patra

Although the Laplace approximation offers a simple route to uncertainty quantification in deep neural networks, its reliance on inverting large Hessian matrices has motivated a range of computationally feasible low-dimensional or sparse approximations. A prominent class of such methods - sub-network Laplace approximations, constructs surrogates by restricting attention to a small subset of parameters. Existing approaches in this family typically rely on diagonal, layer-wise, or other architectural heuristics for subset selection, which ignore cross-parameter interactions and lack formal optimality guarantees. In this paper, we provide a rigorous theoretical analysis of the sub-network Laplace paradigm. We prove that all sub-network Laplace methods systematically underestimate the predictive variance of the full Laplace posterior, and that this bias decreases monotonically as the retained sub-matrix expands. Leveraging this insight, we propose two principled, analytically grounded sub-network Hessian approximations: \textit{Gradient-Laplace} selects parameters with the largest average squared gradients of the model output with respect to the parameters over a reference dataset; while \textit{Greedy-Laplace} iteratively refines this selection by accounting for off-diagonal interactions in the precision matrix. We establish theoretical guarantees characterizing their optimality properties and show that Gradient-Laplace provably outperforms existing heuristic approaches. Extensive numerical studies across diverse settings indicate that these methods perform strongly relative to existing benchmarks.

arXiv

Shot-based quantum encoding: a data-loading paradigm for quantum neural networks

Basil Kyriacou, Viktoria Patapovich, Maniraman Periyasamy, Alexey Melnikov

Efficient data loading remains a bottleneck for near-term quantum machine learning. Existing schemes (angle, amplitude, and basis encoding) either underuse the exponential Hilbert-space capacity or require circuit depths that exceed the coherence budgets of noisy intermediate-scale quantum hardware. We introduce shot-based quantum encoding (SBQE), a data embedding strategy that distributes the hardware's native resource, shots, according to a data-dependent classical distribution over multiple initial quantum states. By treating the shot counts as a learnable degree of freedom, SBQE produces a mixed-state representation whose expectation values are linear in the classical probabilities and can therefore be composed with nonlinear activation functions. We show that SBQE is structurally equivalent to a multilayer perceptron whose weights are realized by quantum circuits, and we describe a hardware-compatible implementation protocol. Benchmarks on three image datasets, with 10 independent initializations per model, show that SBQE achieves 89.1% +- 0.9% test accuracy on Semeion (reducing error by 5.3% relative to amplitude encoding and matching a width-matched classical network), 80.95% +- 0.10% on Fashion MNIST (exceeding amplitude encoding by +2.0% and a linear multilayer perceptron by +1.3%), and 90.25% +- 0.18% on MNIST (exceeding amplitude encoding by 2.1 percentage points and the width-matched classical network by 0.3), all without any data-encoding gates.

arXiv

Adaptively Robust LLM Monitoring via Activation Watermarking

Toluwani Aremu, Daniil Ognev, Samuele Poppi, Nils Lukas

Providers monitor deployed large language models (LLMs) to detect misuse that they cannot prevent. LLM monitoring is deterministic and often openly available, so $\emph{adaptive}$ attackers with a local copy can search offline for prompts that elicit harmful behavior and evade detection. These attacks are especially concerning because providers never observe the misuse and cannot patch their defenses post-hoc. The core challenge is resisting adaptive attackers while preserving detection rates against non-adaptive ones. We propose $\emph{Activation Watermarking}$ (AWM), which randomizes monitoring through limited fine-tuning that aligns the LLM's hidden states with a secret key-derived direction whenever a response violates a policy. Detection is a similarity test on activations the provider already computes, and attackers who know everything but the key must optimize against differently keyed surrogate detectors. At a matched $1\%$ false-positive rate, such surrogate-based attacks evade every evaluated baseline at least $79\%$ of the time, but AWM more than halves this evasion rate. AWM also achieves the lowest evasion rate on three of four non-adaptive jailbreak families at a small drop in utility measured across seven benchmarks. Because random high-dimensional directions are near-orthogonal, assigning one watermark per policy enables attribution. Across 20 monitored policies, AWM identifies the violated policy with $80\%$ accuracy versus a $5\%$ chance baseline. AWM is (i) efficient, (ii) substantially more robust against adaptive attackers than related work and (iii) can attribute which policies were violated.

arXiv

TabPFN Extensions for Interpretable Geotechnical Modelling

Taiga Saito, Yu Otake, Daijiro Mizutani, Stephen Wu

Geotechnical site characterisation relies on sparse, heterogeneous borehole data, where uncertainty quantification and interpretability matter as much as predictive accuracy. We evaluate TabPFN~\citep{Hollmann2025}, a tabular foundation model, and its \texttt{tabpfn-extensions} library on two geotechnical tasks: (1) soil-type classification from N-value and shear-wave velocity data as a controlled illustrative case, and (2) iterative imputation of five mechanical parameters ($s_\mathrm{u}$, $E_{\mathrm{u}}$, ${σ'}_\mathrm{p}$, $C_\mathrm{c}$, $C_\mathrm{v}$) in BM/AirportSoilProperties/2/2025. Without retraining, we apply cosine-similarity analysis to TabPFN embeddings, visualise predictive distributions, and compute SHAP attributions. On the regression benchmark we compare TabPFN with mean imputation, linear regression, random forests, XGBoost, and HBM; introduce a proxy decomposition of predictive uncertainty across context-perturbation classes; and propagate marginal $C_\mathrm{c}$ and ${σ'}_\mathrm{p}$ distributions through a one-dimensional consolidation model to obtain the reliability index $β$ and serviceability exceedance probability $P_\mathrm{f}$. Embeddings exhibit label-consistent Clay/Sand grouping; iterative imputation reduces RMSE for all five targets, with TabPFN lowest on four; SHAP attributions are consistent with the Skempton compression-index correlation and the inverse preconsolidation-pressure-water-content dependence; the within-posterior component is largest in the proxy decomposition. We position the contribution as a worked evaluation workflow that may complement established methods for data-scarce geotechnics, not as algorithmic innovation.

arXiv

SENSE: Efficient EEG-to-Text via Privacy-Preserving Semantic Retrieval

Akshaj Murhekar, Christina Liu, Abhijit Mishra, Shounak Roychowdhury, Jacek Gwizdka

Decoding brain activity into natural language is a major challenge in AI with important applications in assistive communication, neurotechnology, and human-computer interaction. Most existing Brain-Computer Interface (BCI) approaches rely on memory-intensive fine-tuning of Large Language Models (LLMs) or encoder-decoder models on raw EEG signals, resulting in expensive training pipelines, limited accessibility, and potential exposure of sensitive neural data. We introduce SENSE (SEmantic Neural Sparse Extraction), a lightweight and privacy-preserving framework that translates non-invasive electroencephalography (EEG) into text without LLM fine-tuning. SENSE decouples decoding into two stages: on-device semantic retrieval and prompt-based language generation. EEG signals are locally mapped to a discrete textual space to extract a non-sensitive Bag-of-Words (BoW), which conditions an off-the-shelf LLM to synthesize fluent text in a zero-shot manner. The EEG-to-keyword module contains only ~6M parameters and runs fully on-device, ensuring raw neural signals remain local while only abstract semantic cues interact with language models. Evaluated on a 128-channel EEG dataset across six subjects, SENSE matches or surpasses the generative quality of fully fine-tuned baselines such as Thought2Text while substantially reducing computational overhead. By localizing neural decoding and sharing only derived textual cues, SENSE provides a scalable and privacy-aware retrieval-augmented architecture for next-generation BCIs.

arXiv

Transformers Can Learn Rules They've Never Seen: Proof of Computation Beyond Interpolation

Andy Gray

A central question in the debate over large language models is whether transformers can learn rules they have never seen, or whether they can only interpolate: predict new cases from their similarity to training examples. We test this in a controlled setting where interpolation provably fails, so success can only come from computation beyond interpolation. We train small transformers to predict the rollout of a cellular automaton whose update rule is pure XOR, and remove one entry of the rule's truth table from all direct supervision. The missing entry's output is never shown to the model; its only trace is indirect, as wrong values corrupt visible predictions at later timesteps. Because XOR parity flips whenever one input bit is changed, every one-bit neighbour of the missing entry carries the opposite label, and we prove that similarity-based predictors, including nearest-neighbour, kernel, and Gaussian-process methods, are forced to the wrong answer. A two-layer transformer can nevertheless recover the missing entry, and circuit extraction confirms it computes XOR exactly. Ablations show the recovery depends on gradient signal propagating through multi-step prediction, and a second, structurally unrelated benchmark on symbolic operator chains exhibits the same capacity under ordinary autoregressive training. Together with a constructive proof that a standard transformer block can implement exact local Boolean rules, these results provide an existence proof that transformers can learn rule structure not directly observed in training and express it explicitly. This rules out the strongest architectural form of the interpolation-only account, the claim that transformers cannot in principle discover and communicate unseen rules, while leaving open when such behaviour arises in large-scale language training.

arXiv

Persistence Spheres: a Bi-continuous Linear Representation of Measures for Partial Optimal Transport

Matteo Pegoraro

We improve and extend persistence spheres, introduced in~\cite{pegoraro2025persistence}. Persistence spheres map an integrable measure $μ$ on the upper half-plane, including persistence diagrams (PDs) as counting measures, to a function $S(μ)\in C(\mathbb{S}^2)$, and the map is stable with respect to 1-Wasserstein partial transport distance $\mathrm{POT}_1$. Moreover, to the best of our knowledge, persistence spheres are the first explicit representation used in topological machine learning for which continuity of the inverse on the image is established at every compactly supported target. Recent bounded-cardinality bi-Lipschitz embedding results in partial transport spaces, despite being powerful, are not given by the kind of explicit summary map considered here. Our construction is rooted in convex geometry: for positive measures, the defining ReLU integral is the support function of the lift zonoid. Building on~\cite{pegoraro2025persistence}, we refine the definition to better match the $\mathrm{POT}_1$ deletion mechanism, encoding partial transport via a signed diagonal augmentation. In particular, for integrable $μ$, the uniform norm between $S(0)$ and $S(μ)$ depends only on the persistence of $μ$, without any need of ad-hoc re-weightings, reflecting optimal transport to the diagonal at persistence cost. This yields a parameter-free representation at the level of measures (up to numerical discretization), while accommodating future extensions where $μ$ is a smoothed measure derived from PDs (e.g., persistence intensity functions~\citep{wu2024estimation}). Across clustering, regression, and classification tasks involving functional data, time series, graphs, meshes, and point clouds, the updated persistence spheres are competitive and often improve upon persistence images, persistence landscapes, persistence splines, and sliced Wasserstein kernel baselines.

arXiv

The Rise of AI in Weather and Climate Information and its Impact on Global Inequality

Amirpasha Mozaffari, Amanda Duarte, Lina Teckentrup, Stefano Materia, Gina E. C. Charnley, Lluis Palma, Eulalia Baulenas Serra, Dragana Bojovic, Paula Checchia, Aude Carreric, Francisco Doblas-Reyes

AI development's current trajectory risks automating and amplifying the North-South divide in the global climate information system. Frontier models are built almost exclusively in the Global North, and this inequality continues through inputs, processes, and outputs, from biased training data to unrepresentative validation, disproportionately affecting vulnerable regions. Addressing these disparities requires a Climate Digital Public Infrastructure, evaluation metrics centring well-being, and knowledge co-production to foster resilience rather than inequity.

arXiv

MapTab: A Diagnostic Benchmark for Long-Horizon Multi-Criteria Multimodal Reasoning on Heterogeneous Topological Graphs

Ziqiao Shang, Ling-Yue Ge, Zian Xu, Zi-Jian Cheng, Shi-Yu Tian, Zhenyu Huang, Wenbo Fu, Weiming Wu, Yang Chen, Xiangwen Zhang, Yulan Hu, Bin Liu, Lan-Zhe Guo

Systematically evaluating Multimodal Large Language Models (MLLMs) is essential for advancing Artificial General Intelligence (AGI). Yet existing benchmarks remain inadequate for rigorously measuring their reasoning capabilities under multi-criteria constraints. To address this gap, we introduce MapTab, a multimodal benchmark designed to assess holistic multi-criteria reasoning in MLLMs through route-planning tasks. MapTab requires models to perceive and ground visual information from map images while integrating route attributes, such as Time and Price, from structured tables. It covers two scenarios: Metromap, spanning metro networks in 160 cities across 52 countries, and Travelmap, featuring 168 representative tourist attractions from 19 countries. Overall, MapTab includes 328 images, 196,800 route-planning queries, and 3,936 QA queries, incorporating four key criteria: Time, Price, Comfort, and Reliability. Extensive evaluations of 21 representative MLLMs show that current models still struggle with multicriteria multimodal reasoning. Notably, when visual perception is unreliable, multimodal reasoning can even underperform unimodal approaches. MapTab therefore offers a challenging and realistic testbed for systematically evaluating and advancing MLLMs across core perception, integration, numerical comparison, and route planning capabilities.

arXiv

Structurally Separated Uncertainty in Supervised Latent Variable Models

Tanmoy Mukherjee, Marius Kloft, Pierre Marquis, Zied Bouraoui

Predictive uncertainty is commonly decomposed into epistemic and aleatoric components, but standard decompositions often produce strongly correlated estimates because both quantities are derived from the same predictive distribution. We study an alternative design principle, \emph{structural separation}, which assigns epistemic and aleatoric uncertainty to disjoint parameter paths trained with distinct supervision targets: reducible prediction error for epistemic uncertainty and persistent label ambiguity for aleatoric uncertainty. We instantiate this principle in two supervised latent-variable models, a Credal Concept Bottleneck Model and a credal Self-Explaining Neural Network, and prove a gradient-isolation result showing that the two uncertainty heads are not coupled through shared training gradients under the proposed parameterization. Across five ambiguity-aware benchmarks, structural separation substantially reduces epistemic-aleatoric correlation while preserving predictive performance. Further analyses show that aleatoric estimates track annotator- or corpus-derived ambiguity, while epistemic estimates are more sensitive to prediction error and data availability. These results suggest that supervised latent-variable architectures provide a practical route toward uncertainty estimates that are not merely decorrelated, but operationally distinguishable.

arXiv

On the Rademacher Complexity of Graph Neural Networks: Unifying Expressivity and Geometry

Martin Carrasco, Caio F. Deberaldini Netto, Vahan A. Martirosyan, Ehimare Okoyomon, Caterina Graziani

Understanding the interplay between generalization, expressivity, and the geometry of the input space is a central challenge in graph learning. The expressivity of Graph Neural Networks (GNNs) is typically characterized through their correspondence with graph invariants, such as those from the Weisfeiler-Leman (WL) hierarchy. While more expressive GNNs can distinguish a richer set of graphs, they are also associated with weaker generalization guarantees. Previous works have addressed this trade-off using the VC dimension, a purely combinatorial measure, independent of the training data. In this work, we adopt a data-dependent measure of generalization, the empirical Rademacher complexity, and derive tight generalization bounds that jointly consider the expressive power of GNNs and the geometry of the underlying input space. Specifically, any graph invariant that upper-bounds a GNN's expressive power partitions the input space into equivalence classes, and we show that the empirical Rademacher complexity is controlled by the distribution of training samples across these classes. Moving beyond discrete partitions, we incorporate the geometry of the input space and derive covering-number bounds under Lipschitz continuity, showing that the complexity cost can be mitigated when the hypothesis class remains smooth over the data geometry. In addition, we prove that the empirical Rademacher complexity is Lipschitz continuous with respect to the Wasserstein distance between empirical measures supported on different datasets. This yields robustness and generalization guarantees under sampling variability. Importantly, our framework is not restricted to message-passing GNNs or WL, but extends to arbitrary GNN architectures and their associated invariants, providing a step toward a unified theory of GNN generalization.

arXiv

HealthSLM-Bench: Benchmarking Small Language Models for Mobile and Wearable Healthcare Monitoring

Xin Wang, Ting Dang, Xinyu Zhang, Vassilis Kostakos, Michael J. Witbrock, Hong Jia

Mobile and wearable healthcare monitoring play a vital role in facilitating timely interventions, managing chronic health conditions, and ultimately improving individuals' quality of life. Previous studies on large language models (LLMs) have highlighted their impressive generalization abilities and effectiveness in healthcare prediction tasks. However, most LLM-based healthcare solutions are cloud-based, which raises significant privacy concerns and results in increased memory usage and latency. To address these challenges, there is growing interest in compact models, Small Language Models (SLMs), which are lightweight and designed to run locally and efficiently on mobile and wearable devices. Nevertheless, how well these models perform in healthcare prediction remains largely unexplored. We systematically evaluated SLMs on health prediction tasks using zero-shot, few-shot, and instruction fine-tuning approaches, and deployed the best performing fine-tuned SLMs on mobile devices to evaluate their real-world efficiency and predictive performance in practical healthcare scenarios. Our results show that SLMs can achieve performance comparable to LLMs while offering substantial gains in efficiency and privacy. However, challenges remain, particularly in handling class imbalance and few-shot scenarios. These findings highlight SLMs, though imperfect in their current form, as a promising solution for next-generation, privacy-preserving healthcare monitoring.

arXiv

The Advantage of Fine-Grained Training

Davide Pirovano, Federico Milanesio, Michele Caselle, Piero Fariselli, Matteo Osella

In classification problems, models are trained to predict a class label based on the input data features. However, class labels are organized hierarchically in many datasets. While a classification task is often defined at a specific level of this hierarchy, training can utilize a finer granularity of labels. Empirical evidence suggests that such fine-grained training can enhance performance. In this work, we investigate the generality of this observation and explore its underlying causes using both real and synthetic datasets. We show that training on fine-grained labels does not universally improve classification accuracy. Instead, the effectiveness of this strategy depends on the geometric structure of the data and its relations with the label hierarchy. Specifically, we show that the advantage of fine-grained training crucially depends on the degree of alignment between the decision boundaries required for the fine- and coarse-grained tasks, a property that we term boundary redundancy. Additionally, factors such as dataset size and model capacity significantly influence whether fine-grained labels provide a performance benefit. Indeed, we identify a transition, whose location is largely controlled by the degree of overparameterization, separating regimes where fine-grained training improves performance from those where direct coarse-grained training is preferable.

arXiv

ReDiSC: A Reparameterized Masked Diffusion Model for Scalable Node Classification with Structured Predictions

Yule Li, Yifeng Lu, Zhen Wang, Zhewei Wei, Yaliang Li, Bolin Ding

In recent years, graph neural networks (GNN) have achieved unprecedented successes in node classification tasks. Although GNNs inherently encode specific inductive biases (e.g., acting as low-pass or high-pass filters), most existing methods implicitly assume conditional independence among node labels in their optimization objectives. While this assumption is suitable for traditional classification tasks such as image recognition, it contradicts the intuitive observation that node labels in graphs remain correlated, even after conditioning on the graph structure. To make structured predictions for node labels, we propose ReDiSC, namely, Reparameterized masked Diffusion model for Structured node Classification. ReDiSC estimates the joint distribution of node labels using a reparameterized masked diffusion model, which is learned through the variational expectation-maximization (EM) framework. Our theoretical analysis shows the efficiency advantage of ReDiSC in the E-step compared to DPM-SNC, a state-of-the-art model that relies on a manifold-constrained diffusion model in continuous domain. Meanwhile, we explicitly link ReDiSC's M-step objective to popular GNN and label propagation hybrid approaches. Extensive experiments demonstrate that ReDiSC achieves superior or highly competitive performance compared to state-of-the-art GNN, label propagation, and diffusion-based baselines across both homophilic and heterophilic graphs of varying sizes. Notably, ReDiSC scales effectively to large-scale datasets on which previous structured diffusion methods fail due to computational constraints, highlighting its significant practical advantage in structured node classification tasks.

arXiv

Equivariant Eikonal Neural Networks: Grid-Free, Scalable Travel-Time Prediction on Homogeneous Spaces

Alejandro García-Castellanos, David R Wessels, Nicky J. van den Berg, Remco Duits, Daniël M. Pelt, Erik J Bekkers

We introduce Equivariant Neural Eikonal Solvers, a novel framework that integrates Equivariant Neural Fields (ENFs) with Neural Eikonal Solvers. Our approach employs a single neural field where a unified shared backbone is conditioned on signal-specific latent variables - represented as point clouds in a Lie group - to model diverse Eikonal solutions. The ENF integration ensures equivariant mapping from these latent representations to the solution field, delivering three key benefits: enhanced representation efficiency through weight-sharing, robust geometric grounding, and solution steerability. This steerability allows transformations applied to the latent point cloud to induce predictable, geometrically meaningful modifications in the resulting Eikonal solution. By coupling these steerable representations with Physics-Informed Neural Networks (PINNs), our framework accurately models Eikonal travel-time solutions while generalizing to arbitrary Riemannian manifolds with regular group actions. This includes homogeneous spaces such as Euclidean, position-orientation, spherical, and hyperbolic manifolds. We validate our approach through applications in seismic travel-time modeling of 2D, 3D, and spherical benchmark datasets. Experimental results demonstrate superior performance, scalability, adaptability, and user controllability compared to existing Neural Operator-based Eikonal solver methods.

arXiv

Challenges and proposed solutions in modeling multimodal medical data: A systematic review

Maryam Farhadizadeh, Maria Weymann, Michael Blaß, Johann Kraus, Christopher Gundler, Sebastian Walter, Noah Hempen, Hannah Bast, Harald Binder, Nadine Binder

Multimodal data modeling has emerged as a powerful approach in clinical research, enabling the integration of diverse data types such as imaging, genomics, wearable sensors, and electronic health records. Despite its potential to improve diagnostic accuracy and support personalized care, modeling such heterogeneous data presents significant technical challenges. This systematic review synthesizes findings from 69 studies to identify common obstacles, including missing modalities, limited sample sizes, dimensionality imbalance, interpretability issues, and finding the optimal fusion techniques. We highlight recent methodological advances, such as transfer learning, generative models, attention mechanisms, and neural architecture search that offer promising solutions. By mapping current trends and innovations, this review provides a comprehensive overview of the field and offers practical insights to guide future research and development in multimodal modeling for medical applications.

arXiv

Low-Precision Training of Large Language Models: Methods, Challenges, and Opportunities

Zhiwei Hao, Jianyuan Guo, Li Shen, Yong Luo, Han Hu, Guoxia Wang, Dianhai Yu, Yonggang Wen, Dacheng Tao

Large language models (LLMs) have achieved impressive performance across various domains. However, the substantial hardware resources required for their training present a significant barrier to efficiency and scalability. To mitigate this challenge, low-precision training techniques have been widely adopted, leading to notable advancements in training efficiency. Despite these gains, low-precision training involves several components, such as weights, activations, and gradients, each of which can be represented in different numerical formats. The resulting diversity has created a fragmented landscape in low-precision training research, making it difficult for researchers to gain a unified overview of the field. This survey provides a comprehensive review of existing low-precision training methods. To systematically organize these approaches, we categorize them into three primary groups based on their underlying numerical formats, which is a key factor influencing hardware compatibility, computational efficiency, and ease of reference for readers. The categories are (1) fixed-point and integer-based methods, (2) floating-point-based methods, and (3) customized format-based methods. Additionally, we discuss quantization-aware training approaches, which share key similarities with low-precision training during forward propagation. Beyond efficiency, we examine robustness and deployment reliability under low precision. Finally, we highlight several promising research directions to advance this field. A collection of papers discussed in this survey is provided in https://github.com/Hao840/Awesome-Low-Precision-Training.

arXiv

AI Alignment in Medical Imaging: Unveiling Hidden Biases Through Counterfactual Analysis

Haroui Ma, Francesco Quinzan, Theresa Willem, Stefan Bauer

Machine learning (ML) systems for medical imaging have demonstrated remarkable diagnostic capabilities, but their susceptibility to biases poses significant risks, since biases may negatively impact generalization performance. In this paper, we introduce a novel statistical framework to evaluate the dependency of medical imaging ML models on sensitive attributes, such as demographics. Our method leverages the concept of counterfactual invariance, measuring the extent to which a model's predictions remain unchanged under hypothetical changes to sensitive attributes. We present a practical algorithm that combines conditional latent diffusion models with statistical hypothesis testing to identify and quantify such biases without requiring direct access to counterfactual data. Through experiments on synthetic datasets and large-scale real-world medical imaging datasets, including \textsc{cheXpert} and MIMIC-CXR, we demonstrate that our approach aligns closely with counterfactual fairness principles and outperforms standard baselines. This work provides a robust tool to ensure that ML diagnostic systems generalize well, e.g., across demographic groups, offering a critical step towards AI safety in healthcare. Code: https://github.com/Neferpitou3871/AI-Alignment-Medical-Imaging.

arXiv

Optimal Causal Annotations: An Application to Casenotes in Social Services

Ezinne Nwankwo, Lauri Goldkind, Angela Zhou

Problem definition: Estimating causal effects of interventions is central to policy and operations, but outcome data are often missing or costly to obtain. LLMs can provide text annotation at scale but may be subject to unknown bias. When ground-truth outcomes require expensive expert labeling or follow-up, budget limits typically allow only a fraction of the data to be labeled. Motivated by collaboration with a nonprofit conducting street outreach in homelessness services, whose most interesting outcomes are in unstructured casenotes, we ask: which observations should be selected for labeling under a fixed budget? Methodology/results: Our method optimizes annotation probabilities to minimize the variance of average treatment effect estimation. We derive a closed-form solution and establish that a feasible two-batch estimator achieves the best possible asymptotic variance. On simulated and real-world datasets, our method achieves lower MSE than random sampling and 43%-91% reductions in labels and costs for the same interval widths. In a case study classifying progress towards housing applications from casenotes with LLMs, we find out-of-the-box LLMs under-recognize client progress, highlighting the importance of grounding LLM judgment. Our estimates indicate that 8.6% of clients improve 2-year housing outcomes due to more street outreach in the first six months, and increasing outreach from the first to third quartile increases maximum progress towards a housing application by an estimated half-step. Managerial implications: LLM-as-a-judge may threaten estimation validity; our method enables valid causal estimation with a limited annotation budget. Evaluating causal impacts on intermediate nonprofit outcomes can inform operations: returns to outreach appear concave, so expanding the extensive margin of who receives outreach may be resource-efficient.

arXiv

Learning Controlled Stochastic Differential Equations

Luc Brogat-Motte, Riccardo Bonalli, Alessandro Rudi

We study the problem of learning controlled stochastic differential equations (SDEs) \[ dX_t = b(t,X_t,u_t)\,dt + σ(t,X_t,u_t)\,dW_t, \] whose drift and diffusion depend nonlinearly on time, state, and control values. From trajectory data, we aim to estimate coefficients whose induced density flows reproduce those of the observed dynamics. The data consist of several controls sampled from a finite-dimensional family and, for each control, multiple independent trajectories observed at discrete times over a fixed horizon. The controls are observed inputs, not learner-selected decisions. We propose a kernel method for multidimensional nonlinear controlled SDEs. The method estimates the density flow for each sampled control, then fits the drift $b$ and diffusion matrix $a=σσ^\top$ by matching the estimated flows through the Fokker-Planck equation. Under Sobolev regularity assumptions, we prove finite-sample bounds on the error between the density flows of the learned and true SDEs, measured in $L^2$ over controls, time, and state. The bounds quantify how the error decreases with the number $K$ of sampled controls, with rates determined by the state and control-parametrization dimensions and Sobolev regularity. We further derive uniform-in-control guarantees and CVaR-type bounds for tail-sensitive quantities. Numerical experiments illustrate the method, and an open-source Python implementation is provided.

Keep browsing

More papers from top AI conferences

30+ top conferences, by conference, year, and topic.