arXiv subjects

Patrick Langer

Publications and source records attributed to Patrick Langer.

12 recordsLinked to original sources

arXiv2026

TimeRLM: Recursive Language Models Enable Precise Anomaly Localization in Long-Context Time-Series

Precise anomaly localization over long-context time series is a crucial task in monitoring applications across clinical care, industrial operations, financial services, and logistics, where brief evidence may hide inside long spans of high-frequency data. Time-Series Language Models (TSLMs) are able to ingest time series data and verbalize findings on anomalies in natural language; however, recent benchmarks report a decrease in retrieval performance at long contexts, mirroring failure modes in text, vision, and audio. In the text domain, Recursive Language Models (RLMs) can recover much of this lost performance by keeping context external to the large language model (LLM), allowing the model to query it through code. We present TimeRLM, an RLM formulation for time-series that sequentially manipulates the signal using code and vision capabilities. We further introduce AnomalyXL, a synthetic long-context anomaly localization benchmark with programmatically injected anomalies that require precise retrieval. We implement five different task categories and two variants: AnomalyXL-MCQ and AnomalyXL-Localize. TimeRLM outperforms every evaluated TSLM and single-pass baseline on four of the five AnomalyXL-Localize tasks, reaching 0.682 IoU on localization and 0.745 on classify-with-evidence, versus at most 0.329 and 0.072 across all baselines. We post-train TimeRLM using reinforcement learning. The resulting model further improves performance and requires approximately one-third as many agent interaction turns as its untrained base model to produce a final answer. On unseen real-world ECG, sleep and software observability recordings, the post-trained TimeRLM retains or improves performance, surpassing TSLMs despite being trained exclusively on synthetic data. Our findings suggest recursive interaction with time-series is an effective approach for long-horizon retrieval.

cs.LG↗

arXiv2026

ARA: Agentic Reproducibility Assessment For Scalable Support Of Scientific Peer-Review

Scientific peer review increasingly struggles to assess reproducibility at the scale and complexity of modern research output. Evaluating reproducibility requires reconstructing experimental dependencies, methodological choices, data flows, and result-generating procedures, which often exceeds what human reviewers can provide. Agentic Reproducibility Assessment (ARA) formalizes reproducibility assessment as a structured reasoning task over scientific documents. Given a paper, ARA extracts a directed workflow graph linking sources, methods, experiments, and outputs, then evaluates its reconstructability using structural and content-based scores for reproducibility assessments. Experiments on 213 ReScience C articles - the largest cross-domain benchmark of human-validated computational reproducibility studies considered to date - demonstrate ARA's generalizability and consistent workflow reconstruction and assessment across LLMs, model temperatures, and scientific domains. ARA achieves ~61% accuracy on three benchmarks, and the highest accuracy reported on ReproBench (60.71% vs. 36.84%) and GoldStandardDB (61.68% vs. 43.56%), highlighting its potential to complement human review at scale and enabling next-generation peer review. Code and Data available: https://github.com/AndresLaverdeMarin/agentic_reproducibility_assessment.

cs.DL↗

arXiv2026

New Bounds for Zarankiewicz Numbers via Reinforced LLM Evolutionary Search

The Zarankiewicz number $\textbf{Z}(m, n, s, t)$ is the maximum number of edges in a bipartite graph $G_{m, n}$ such that there is no complete $K_{s, t}$ bipartite subgraph. We determine for the first time the exact values of three Zarankiewicz numbers: $\textbf{Z}(11, 21, 3, 3)=116$, $\textbf{Z}(11, 22, 3, 3)=121$, and $\textbf{Z}(12, 22, 3, 3)=132$. We further establish lower bounds for 41 more Zarankiewicz numbers, including several that are within one edge of the best known upper bound, and we match the established value in four more closed cases. Our results are obtained using OpenEvolve, an open-source evolutionary algorithm based on Large Language Models (LLMs) that iteratively improves algorithms for generating mathematical constructions by optimizing a reward signal which we tailored for this specific problem. These findings provide new extremal graph constructions and demonstrate the potential of LLM-guided evolutionary search to contribute to mathematical research. In addition to presenting the resulting constructions, we report the generation algorithms produced, describe the relevant implementation details, and provide our computational costs. Our costs are remarkably low, at less than \$30 for each Zarankiewicz parameter combination, showing that LLM-guided evolutionary search can be an inexpensive, reproducible, and accessible tool for discovering new combinatorial constructions.

cs.AI↗

arXiv2026

Patient-Level Multimodal Question Answering from Multi-Site Auscultation Recordings

Auscultation is a vital diagnostic tool, yet its utility is often limited by subjective interpretation. While general-purpose Audio-Language Models (ALMs) excel in general domains, they struggle with the nuances of physiological signals. We propose a framework that aligns multi-site auscultation recordings directly with a frozen Large Language Model (LLM) embedding space via gated cross-attention. By leveraging the LLM's latent world knowledge, our approach moves beyond isolated classification toward holistic, patient-level assessment. On the CaReSound benchmark, our model achieves a state-of-the-art 0.865 F1-macro and 0.952 BERTScore. We demonstrate that lightweight, domain-specific encoders rival large-scale ALMs and that multi-site aggregation provides spatial redundancy that mitigates temporal truncation. This alignment of medical acoustics with text foundations offers a scalable path for bridging signal processing and clinical assessment.

cs.SD↗

arXiv2026

How Well Do Multimodal Models Reason on ECG Signals?

While multimodal large language models offer a promising solution to the "black box" nature of health AI by generating interpretable reasoning traces, verifying the validity of these traces remains a critical challenge. Existing evaluation methods are either unscalable, relying on manual clinician review, or superficial, utilizing proxy metrics (e.g. QA) that fail to capture the semantic correctness of clinical logic. In this work, we introduce a reproducible framework for evaluating reasoning in ECG signals. We propose decomposing reasoning into two distinct, components: (i) Perception, the accurate identification of patterns within the raw signal, and (ii) Deduction, the logical application of domain knowledge to those patterns. To evaluate Perception, we employ an agentic framework that generates code to empirically verify the temporal structures described in the reasoning trace. To evaluate Deduction, we measure the alignment of the model's logic against a structured database of established clinical criteria in a retrieval-based approach. This dual-verification method enables the scalable assessment of "true" reasoning capabilities.

cs.AI↗

arXiv2026

TS-Haystack: A Multi-Task Retrieval Benchmark for Long-Context Time-Series Reasoning

Time Series Language Models (TSLMs) promise reasoning over real-world temporal data, but their ability to retrieve and reason over long time-series remains largely untested. We introduce TS-Haystack, a multi-domain retrieval benchmark with ten event-grounded question-answering tasks over contexts from 100 seconds to 24 hours, spanning direct retrieval, temporal reasoning, multi-step reasoning, and contextual anomaly detection. Existing TSLMs exhibit severe long-context degradation: accuracy declines with context length, direct-tokenization models run out of memory beyond 100 seconds on high-rate signals, and time-interval-grounded tasks collapse toward near-zero accuracy when increasing the time-series lengths, aligning with existing literature on text and multi-modal long context retrieval. An agentic retrieval framework using specialized time-series classifier tools matches or outperforms SoTA TSLMs on 9 of 10 tasks, highlighting agentic retrieval as a promising approach for long-context TSLMs.

cs.LG↗

arXiv2026

EvoMorph: Counterfactual Explanations for Continuous Time-Series Extrinsic Regression Applied to Photoplethysmography

Wearable devices enable continuous, population-scale monitoring of physiological signals, such as photoplethysmography (PPG), creating new opportunities for data-driven clinical assessment. Time-series extrinsic regression (TSER) models increasingly leverage PPG signals to estimate clinically relevant outcomes, including heart rate, respiratory rate, and oxygen saturation. For clinical reasoning and trust, however, single point estimates alone are insufficient: clinicians must also understand whether predictions are stable under physiologically plausible variations and to what extent realistic, attainable changes in physiological signals would meaningfully alter a model's prediction. Counterfactual explanations (CFE) address these "what-if" questions, yet existing time series CFE generation methods are largely restricted to classification, overlook waveform morphology, and often produce physiologically implausible signals, limiting their applicability to continuous biomedical time series. To address these limitations, we introduce EvoMorph, a multi-objective evolutionary framework for generating physiologically plausible and diverse CFE for TSER applications. EvoMorph optimizes morphology-aware objectives defined on interpretable signal descriptors and applies transformations to preserve the waveform structure. We evaluated EvoMorph on three PPG datasets (heart rate, respiratory rate, and oxygen saturation) against a nearest-unlike-neighbor baseline. In addition, in a case study, we evaluated EvoMorph as a tool for uncertainty quantification by relating counterfactual sensitivity to bootstrap-ensemble uncertainty and data-density measures. Overall, EvoMorph enables the generation of physiologically-aware counterfactuals for continuous biomedical signals and supports uncertainty-aware interpretability, advancing trustworthy model analysis for clinical time-series applications.

cs.LG↗

arXiv2025

OpenTSLM: Time-Series Language Models for Reasoning over Multivariate Medical Text- and Time-Series Data

LLMs have emerged as powerful tools for interpreting multimodal data. In medicine, they hold particular promise for synthesizing large volumes of clinical information into actionable insights and digital health applications. Yet, a major limitation remains their inability to handle time series. To overcome this gap, we present OpenTSLM, a family of Time Series Language Models (TSLMs) created by integrating time series as a native modality to pretrained LLMs, enabling reasoning over multiple time series of any length. We investigate two architectures for OpenTSLM. The first, OpenTSLM-SoftPrompt, models time series implicitly by concatenating learnable time series tokens with text tokens via soft prompting. Although parameter-efficient, we hypothesize that explicit time series modeling scales better and outperforms implicit approaches. We thus introduce OpenTSLM-Flamingo, which integrates time series with text via cross-attention. We benchmark both variants against baselines that treat time series as text tokens or plots, across a suite of text-time-series Chain-of-Thought (CoT) reasoning tasks. We introduce three datasets: HAR-CoT, Sleep-CoT, and ECG-QA-CoT. Across all, OpenTSLM models outperform baselines, reaching 69.9 F1 in sleep staging and 65.4 in HAR, compared to 9.05 and 52.2 for finetuned text-only models. Notably, even 1B-parameter OpenTSLM models surpass GPT-4o (15.47 and 2.95). OpenTSLM-Flamingo matches OpenTSLM-SoftPrompt in performance and outperforms on longer sequences, while maintaining stable memory requirements. By contrast, SoftPrompt grows exponentially in memory with sequence length, requiring around 110 GB compared to 40 GB VRAM when training on ECG-QA with LLaMA-3B. Expert reviews by clinicians find strong reasoning capabilities exhibited by OpenTSLMs on ECG-QA. To facilitate further research, we provide all code, datasets, and models open-source.

cs.LG↗

arXiv2025

Digital Scale: Open-Source On-Device BMI Estimation from Smartphone Camera Images Trained on a Large-Scale Real-World Dataset

Estimating Body Mass Index (BMI) from camera images with machine learning models enables rapid weight assessment when traditional methods are unavailable or impractical, such as in telehealth or emergency scenarios. Existing computer vision approaches have been limited to datasets of up to 14,500 images. In this study, we present a deep learning-based BMI estimation method trained on our WayBED dataset, a large proprietary collection of 84,963 smartphone images from 25,353 individuals. We introduce an automatic filtering method that uses posture clustering and person detection to curate the dataset by removing low-quality images, such as those with atypical postures or incomplete views. This process retained 71,322 high-quality images suitable for training. We achieve a Mean Absolute Percentage Error (MAPE) of 7.9% on our hold-out test set (WayBED data) using full-body images, the lowest value in the published literature to the best of our knowledge. Further, we achieve a MAPE of 13% on the completely unseen~(during training) VisualBodyToBMI dataset, comparable with state-of-the-art approaches trained on it, demonstrating robust generalization. Lastly, we fine-tune our model on VisualBodyToBMI and achieve a MAPE of 8.56%, the lowest reported value on this dataset so far. We deploy the full pipeline, including image filtering and BMI estimation, on Android devices using the CLAID framework. We release our complete code for model training, filtering, and the CLAID package for mobile deployment as open-source contributions.

cs.CV↗

arXiv2024

Comparative Efficacy of Commercial Wearables for Circadian Rhythm Home Monitoring from Activity, Heart Rate, and Core Body Temperature

Circadian rhythms govern biological patterns that follow a 24-hour cycle. Dysfunctions in circadian rhythms can contribute to various health problems, such as sleep disorders. Current circadian rhythm assessment methods, often invasive or subjective, limit circadian rhythm monitoring to laboratories. Hence, this study aims to investigate scalable consumer-centric wearables for circadian rhythm monitoring outside traditional laboratories. In a two-week longitudinal study conducted in real-world settings, 36 participants wore an Actigraph, a smartwatch, and a core body temperature sensor to collect activity, temperature, and heart rate data. We evaluated circadian rhythms calculated from commercial wearables by comparing them with circadian rhythm reference measures, i.e., Actigraph activities and chronotype questionnaire scores. The circadian rhythm metric acrophases, determined from commercial wearables using activity, heart rate, and temperature data, significantly correlated with the acrophase derived from Actigraph activities (r=0.96, r=0.87, r=0.79; all p<0.001) and chronotype questionnaire (r=-0.66, r=-0.73, r=-0.61; all p<0.001). The acrophases obtained concurrently from consumer sensors significantly predicted the chronotype (R2=0.64; p<0.001). Our study validates commercial sensors for circadian rhythm assessment, highlighting their potential to support maintaining healthy rhythms and provide scalable and timely health monitoring in real-life scenarios.

eess.SP↗

arXiv2023

CLAID: Closing the Loop on AI & Data Collection -- A Cross-Platform Transparent Computing Middleware Framework for Smart Edge-Cloud and Digital Biomarker Applications

The increasing number of edge devices with enhanced sensing capabilities, such as smartphones, wearables, and IoT devices equipped with sensors, holds the potential for innovative smart-edge applications in healthcare. These devices generate vast amounts of multimodal data, enabling the implementation of digital biomarkers which can be leveraged by machine learning solutions to derive insights, predict health risks, and allow personalized interventions. Training these models requires collecting data from edge devices and aggregating it in the cloud. To validate and verify those models, it is essential to utilize them in real-world scenarios and subject them to testing using data from diverse cohorts. Since some models are too computationally expensive to be run on edge devices directly, a collaborative framework between the edge and cloud becomes necessary. In this paper, we present CLAID, an open-source cross-platform middleware framework based on transparent computing compatible with Android, iOS, WearOS, Linux, macOS, and Windows. CLAID enables logical integration of devices running different operating systems into an edge-cloud system, facilitating communication and offloading between them, with bindings available in different programming languages. We provide Modules for data collection from various sensors as well as for the deployment of machine-learning models. Furthermore, we propose a novel methodology, "ML-Model in the Loop" for verifying deployed machine learning models, which helps to analyze problems that may occur during the migration of models from cloud to edge devices. We verify our framework in three different experiments and achieve 100% sampling coverage for data collection across different sensors as well as an equal performance of a cough detection model deployed on both Android and iOS devices. We evaluate the memory and battery consumption of our framework.

cs.DC↗

arXiv2022

Efficient and Robust Semantic Mapping for Indoor Environments

A key proficiency an autonomous mobile robot must have to perform high-level tasks is a strong understanding of its environment. This involves information about what types of objects are present, where they are, what their spatial extend is, and how they can be reached, i.e., information about free space is also crucial. Semantic maps are a powerful instrument providing such information. However, applying semantic segmentation and building 3D maps with high spatial resolution is challenging given limited resources on mobile robots. In this paper, we incorporate semantic information into efficient occupancy normal distribution transform (NDT) maps to enable real-time semantic mapping on mobile robots. On the publicly available dataset Hypersim, we show that, due to their sub-voxel accuracy, semantic NDT maps are superior to other approaches. We compare them to the recent state-of-the-art approach based on voxels and semantic Bayesian spatial kernel inference~(S-BKI) and to an optimized version of it derived in this paper. The proposed semantic NDT maps can represent semantics to the same level of detail, while mapping is 2.7 to 17.5 times faster. For the same grid resolution, they perform significantly better, while mapping is up to more than 5 times faster. Finally, we prove the real-world applicability of semantic NDT maps with qualitative results in a domestic application.

cs.RO↗