Xingyao Wang

Jiaxin Pei, Aparna Ananthasubramaniam, Xingyao Wang et al. · berkeley, stanford

We present POTATO, the Portable text annotation tool, a free, fully open-sourced annotation system that 1) supports labeling many types of text and multimodal data; 2) offers easy-to-configure features to maximize the productivity of both deployers and annotators (convenient templates for common ML/NLP tasks, active learning, keypress shortcuts, keyword highlights, tooltips); and 3) supports a high degree of customization (editable UI, inserting pre-screening questions, attention and qualification tests). Experiments over two annotation tasks suggest that POTATO improves labeling speed through its specially-designed productivity features, especially for long documents and complex tasks. POTATO is available at https://github.com/davidjurgens/potato and will continue to be updated.

Xingyao Wang, Boxuan Li, Yufan Song et al. · berkeley, cmu

Software is one of the most powerful tools that we humans have at our disposal; it allows a skilled programmer to interact with the world in complex and profound ways. At the same time, thanks to improvements in large language models (LLMs), there has also been a rapid development in AI agents that interact with and affect change in their surrounding environments. In this paper, we introduce OpenHands (f.k.a. OpenDevin), a platform for the development of powerful and flexible AI agents that interact with the world in similar ways to those of a human developer: by writing code, interacting with a command line, and browsing the web. We describe how the platform allows for the implementation of new agents, safe interaction with sandboxed environments for code execution, coordination between multiple agents, and incorporation of evaluation benchmarks. Based on our currently incorporated benchmarks, we perform an evaluation of agents over 15 challenging tasks, including software engineering (e.g., SWE-BENCH) and web browsing (e.g., WEBARENA), among others. Released under the permissive MIT license, OpenHands is a community project spanning academia and industry with more than 2.1K contributions from over 188 contributors.

Xingyao Wang, Zihan Wang, Jiateng Liu et al.

To solve complex tasks, large language models (LLMs) often require multiple rounds of interactions with the user, sometimes assisted by external tools. However, current evaluation protocols often emphasize benchmark performance with single-turn exchanges, neglecting the nuanced interactions among the user, LLMs, and external tools, while also underestimating the importance of natural language feedback from users. These oversights contribute to discrepancies between research benchmark evaluations and real-world use cases. We introduce MINT, a benchmark that evaluates LLMs' ability to solve tasks with multi-turn interactions by (1) using tools and (2) leveraging natural language feedback. To ensure reproducibility, we provide an evaluation framework where LLMs can access tools by executing Python code and receive users' natural language feedback simulated by GPT-4. We repurpose a diverse set of established evaluation datasets focusing on reasoning, coding, and decision-making and carefully curate them into a compact subset for efficient evaluation. Our analysis of 20 open- and closed-source LLMs offers intriguing findings. (a) LLMs generally benefit from tools and language feedback, with performance gains (absolute, same below) of 1-8% for each turn of tool use and 2-17% with natural language feedback. (b) Better single-turn performance does not guarantee better multi-turn performance. (c) Surprisingly, on the LLMs evaluated, supervised instruction-finetuning (SIFT) and reinforcement learning from human feedback (RLHF) generally hurt multi-turn capabilities. We expect MINT can help measure progress and incentivize research in improving LLMs' capabilities in multi-turn interactions, especially for open-source communities where multi-turn human evaluation can be less accessible compared to commercial LLMs with a larger user base.

Lifan Yuan, Yangyi Chen, Xingyao Wang et al.

Large language models (LLMs) are often augmented with tools to solve complex tasks. By generating code snippets and executing them through task-specific Application Programming Interfaces (APIs), they can offload certain functions to dedicated external modules, such as image encoding and performing calculations. However, most existing approaches to augment LLMs with tools are constrained by general-purpose APIs and lack the flexibility for tailoring them to specific tasks. In this work, we present CRAFT, a general tool creation and retrieval framework for LLMs. It creates toolsets specifically curated for the tasks and equips LLMs with a component that retrieves tools from these sets to enhance their capability to solve complex tasks. For each task, we collect specific code solutions by prompting GPT-4 to solve the training examples. Following a validation step ensuring the correctness, these solutions are abstracted into code snippets to enhance reusability, and deduplicated for higher quality. At inference time, the language model retrieves snippets from the toolsets and then executes them or generates the output conditioning on the retrieved snippets. Our method is designed to be flexible and offers a plug-and-play approach to adapt off-the-shelf LLMs to unseen domains and modalities, without any finetuning. Experiments on vision-language, tabular processing, and mathematical reasoning tasks show that our approach achieves substantial improvements compared to strong baselines. In addition, our in-depth analysis reveals that: (1) consistent performance improvement can be achieved by scaling up the number of tools and the capability of the backbone models; (2) each component of our approach contributes to the performance gains; (3) the created tools are well-structured and reliable with low complexity and atomicity. The code is available at https://github.com/lifan-yuan/CRAFT.

Yangyi Chen, Xingyao Wang, Manling Li et al.

State-of-the-art vision-language models (VLMs) still have limited performance in structural knowledge extraction, such as relations between objects. In this work, we present ViStruct, a training framework to learn VLMs for effective visual structural knowledge extraction. Two novel designs are incorporated. First, we propose to leverage the inherent structure of programming language to depict visual structural information. This approach enables explicit and consistent representation of visual structural information of multiple granularities, such as concepts, relations, and events, in a well-organized structured format. Second, we introduce curriculum-based learning for VLMs to progressively comprehend visual structures, from fundamental visual concepts to intricate event structures. Our intuition is that lower-level knowledge may contribute to complex visual structure understanding. Furthermore, we compile and release a collection of datasets tailored for visual structural knowledge extraction. We adopt a weakly-supervised approach to directly generate visual event structures from captions for ViStruct training, capitalizing on abundant image-caption pairs from the web. In experiments, we evaluate ViStruct on visual structure prediction tasks, demonstrating its effectiveness in improving the understanding of visual structures. The code is public at \url{https://github.com/Yangyi-Chen/vi-struct}.

Hanning Zhang, Shizhe Diao, Yong Lin et al.

Large language models (LLMs) have revolutionized numerous domains with their impressive performance but still face their challenges. A predominant issue is the propensity for these models to generate non-existent facts, a concern termed hallucination. Our research is motivated by the observation that previous instruction tuning methods force the model to complete a sentence no matter whether the model knows the knowledge or not. When the question is out of the parametric knowledge, it will try to make up something and fail to indicate when it lacks knowledge. In this paper, we present a new approach called Refusal-Aware Instruction Tuning (R-Tuning). This approach is formalized by first identifying the disparity in knowledge encompassed by pre-trained parameters compared to that of instruction tuning data. Then, we construct the refusal-aware data based on the knowledge intersection, to tune LLMs to refrain from responding to questions beyond its parametric knowledge. Experimental results demonstrate R-Tuning effectively improves a model's ability to answer known questions and refrain from answering unknown questions. Furthermore, when tested on out-of-domain datasets, the refusal ability was found to be a meta-skill that could be generalized to other tasks. Further analysis surprisingly finds that learning the uncertainty results in better calibration and an improved ability to estimate the uncertainty than uncertainty-based testing. Our code is available at https://github.com/shizhediao/R-Tuning.

Yuxiang Lin, Zihan Wang, Mengyang Liu et al.

While agents are increasingly spending more resources, today agent cost is mostly measured only after execution. A Budget-Aware Agent (BAGEN) should treat budget as an active control signal, rather than a passive cost metric. We first systematically define budget estimation as internal budgets (from agent computation) and external budgets (from agent actions). We then formalize budget-awareness as progressive interval estimation: at each step of a plan, an agent should predict an upper and lower bound on remaining budget, and alert when completion is unlikely. Scoring with a rollout-replay protocol, we find consistent failure patterns on four environments and five frontier agents: (1) strong agents do not necessarily have strong budget-awareness, with correlation r=0.35. (2) frontier models are consistently over-optimistic, continue spending on tasks that are unlikely to succeed, instead of alerting the user early. (3) budget-aware signal is actionable and trainable. Early stop saves 28-64% tokens on failed trajectories, and SFT+RL strengthens early stop and alert behavior. (4) precise interval calibration remains challenging, with interval coverage capping at 47% after SFT+RL. Project page: https://ragen-ai.github.io/bagen/

Yangyi Chen, Xingyao Wang, Hao Peng et al.

We present SOLO, a single transformer for Scalable visiOn-Language mOdeling. Current large vision-language models (LVLMs) such as LLaVA mostly employ heterogeneous architectures that connect pre-trained visual encoders with large language models (LLMs) to facilitate visual recognition and complex reasoning. Although achieving remarkable performance with relatively lightweight training, we identify four primary scalability limitations: (1) The visual capacity is constrained by pre-trained visual encoders, which are typically an order of magnitude smaller than LLMs. (2) The heterogeneous architecture complicates the use of established hardware and software infrastructure. (3) Study of scaling laws on such architecture must consider three separate components - visual encoder, connector, and LLMs, which complicates the analysis. (4) The use of existing visual encoders typically requires following a pre-defined specification of image inputs pre-processing, for example, by reshaping inputs to fixed-resolution square images, which presents difficulties in processing and training on high-resolution images or those with unusual aspect ratio. A unified single Transformer architecture, like SOLO, effectively addresses these scalability concerns in LVLMs; however, its limited adoption in the modern context likely stems from the absence of reliable training recipes that balance both modalities and ensure stable training for billion-scale models. In this paper, we introduce the first open-source training recipe for developing SOLO, an open-source 7B LVLM using moderate academic resources. The training recipe involves initializing from LLMs, sequential pre-training on ImageNet and web-scale data, and instruction fine-tuning on our curated high-quality datasets. On extensive evaluation, SOLO demonstrates performance comparable to LLaVA-v1.5-7B, particularly excelling in visual mathematical reasoning.

Xuehang Guo, Zora Zhiruo Wang, Qingyun Wang et al.

Large language models (LLMs) have enabled powerful software engineering (SE) agents capable of navigating complex codebases and resolving real-world issues. However, these agents remain fundamentally episodic: they fail to retain, refine, and reuse experiences across tasks, repeatedly reconstructing context from scratch and reproducing similar mistakes. Even with memory support, they offer no remedy for the absence of a principled, task-agnostic \textit{memory utility}, making them difficult to evaluate rigorously or generalize across agents and settings. To tackle these limitations, we introduce \ours, a closed-loop framework for memory augmentation in SE agents. \ours grounds memory utility in \textit{validated downstream impact}, establishing utility as both a task-agnostic \textbf{evaluation benchmark} and an annotation-free \textbf{optimization signal}. Through complementary evaluation on \textit{single-episode} and \textit{cross-episode} memory augmentation, results demonstrate that \ours consistently improves SE agents across settings, achieving absolute gains of up to $\uparrow5.25\%$ in success rate and $\uparrow4.63\%$ in resolve efficiency, while substantially reducing computational cost by $\geq9.79\%$. Our project page: \href{https://xhguo7.github.io/MemOp/}{https://xhguo7.github.io/MemOp/}.

Hongxiang Gao, Xingyao Wang, Zhenghua Chen et al.

Electrocardiogram (ECG) monitoring is one of the most powerful technique of cardiovascular disease (CVD) early identification, and the introduction of intelligent wearable ECG devices has enabled daily monitoring. However, due to the need for professional expertise in the ECGs interpretation, general public access has once again been restricted, prompting the need for the development of advanced diagnostic algorithms. Classic rule-based algorithms are now completely outperformed by deep learning based methods. But the advancement of smart diagnostic algorithms is hampered by issues like small dataset, inconsistent data labeling, inefficient use of local and global ECG information, memory and inference time consuming deployment of multiple models, and lack of information transfer between tasks. We propose a multi-resolution model that can sustain high-resolution low-level semantic information throughout, with the help of the development of low-resolution high-level semantic information, by capitalizing on both local morphological information and global rhythm information. From the perspective of effective data leverage and inter-task knowledge transfer, we develop a parameter isolation based ECG continual learning (ECG-CL) approach. We evaluated our model's performance on four open-access datasets by designing segmentation-to-classification for cross-domain incremental learning, minority-to-majority class for category incremental learning, and small-to-large sample for task incremental learning. Our approach is shown to successfully extract informative morphological and rhythmic features from ECG segmentation, leading to higher quality classification results. From the perspective of intelligent wearable applications, the possibility of a comprehensive ECG interpretation algorithm based on single-lead ECGs is also confirmed.

Xingyao Wang, Sha Li, Heng Ji

Large Language Model (LLM) trained on a mixture of text and code has demonstrated impressive capability in translating natural language (NL) into structured code. We observe that semantic structures can be conveniently translated into code and propose Code4Struct to leverage such text-to-structure translation capability to tackle structured prediction tasks. As a case study, we formulate Event Argument Extraction (EAE) as converting text into event-argument structures that can be represented as a class object using code. This alignment between structures and code enables us to take advantage of Programming Language (PL) features such as inheritance and type annotation to introduce external knowledge or add constraints. We show that, with sufficient in-context examples, formulating EAE as a code generation problem is advantageous over using variants of text-based prompts. Despite only using 20 training event instances for each event type, Code4Struct is comparable to supervised models trained on 4,202 instances and outperforms current state-of-the-art (SOTA) trained on 20-shot data by 29.5% absolute F1. Code4Struct can use 10-shot training data from a sibling event type to predict arguments for zero-resource event types and outperforms the zero-shot baseline by 12% absolute F1.

Yangyi Chen, Xingyao Wang, Heng Ji

Pre-trained language models (PLMs) serve as backbones for various real-world systems. For high-stake applications, it's equally essential to have reasonable confidence estimations in predictions. While the vanilla confidence scores of PLMs can already be effectively utilized, PLMs consistently become overconfident in their wrong predictions, which is not desirable in practice. Previous work shows that introducing an extra calibration task can mitigate this issue. The basic idea involves acquiring additional data to train models in predicting the confidence of their initial predictions. However, it only demonstrates the feasibility of this kind of method, assuming that there are abundant extra available samples for the introduced calibration task. In this work, we consider the practical scenario that we need to effectively utilize training samples to make PLMs both task-solvers and self-calibrators. Three challenges are presented, including limited training samples, data imbalance, and distribution shifts. We first conduct pilot experiments to quantify various decisive factors in the calibration task. Based on the empirical analysis results, we propose a training algorithm LM-TOAST to tackle the challenges. Experimental results show that LM-TOAST can effectively utilize the training data to make PLMs have reasonable confidence estimations while maintaining the original task performance. Further, we consider three downstream applications, namely selective classification, adversarial defense, and model cascading, to show the practical usefulness of LM-TOAST. The code will be made public at \url{https://github.com/Yangyi-Chen/LM-TOAST}.

Xingyao Wang, Valerie Chen, Heng Ji et al. · cmu

Academic benchmarks for coding agents tend to reward autonomous task completion, measured by verifiable rewards such as unit-test success. In contrast, real-world coding agents operate with humans in the loop, where success signals are typically noisy, delayed, and sparse. How can we bridge this gap? In this paper, we propose a process to learn a "critic" model from sparse and noisy interaction data, which can then be used both as a reward model for either RL-based training or inference-time scaling. Specifically, we introduce Critic Rubrics, a rubric-based supervision framework with 24 behavioral features that can be derived from human-agent interaction traces alone. Using a semi-supervised objective, we can then jointly predict these rubrics and sparse human feedback (when present). In experiments, we demonstrate that, despite being trained primarily from trace-observable rubrics and sparse real-world outcome proxies, these critics improve best-of-N reranking on SWE-bench (Best@8 +15.9 over Random@8 over the rerankable subset of trajectories), enable early stopping (+17.7 with 83% fewer attempts), and support training-time data curation via critic-selected trajectories.

Xingyao Wang, Yuwen Li, Hongxiang Gao et al.

Precise segmentation is a vital first step to analyze semantic information of cardiac cycle and capture anomaly with cardiovascular signals. However, in the field of deep semantic segmentation, inference is often unilaterally confounded by the individual attribute of data. Towards cardiovascular signals, quasi-periodicity is the essential characteristic to be learned, regarded as the synthesize of the attributes of morphology (Am) and rhythm (Ar). Our key insight is to suppress the over-dependence on Am or Ar while the generation process of deep representations. To address this issue, we establish a structural causal model as the foundation to customize the intervention approaches on Am and Ar, respectively. In this paper, we propose contrastive causal intervention (CCI) to form a novel training paradigm under a frame-level contrastive framework. The intervention can eliminate the implicit statistical bias brought by the single attribute and lead to more objective representations. We conduct comprehensive experiments with the controlled condition for QRS location and heart sound segmentation. The final results indicate that our approach can evidently improve the performance by up to 0.41% for QRS location and 2.73% for heart sound segmentation. The efficiency of the proposed method is generalized to multiple databases and noisy signals.

Genglin Liu, Xingyao Wang, Lifan Yuan et al.

Can large language models (LLMs) express their uncertainty in situations where they lack sufficient parametric knowledge to generate reasonable responses? This work aims to systematically investigate LLMs' behaviors in such situations, emphasizing the trade-off between honesty and helpfulness. To tackle the challenge of precisely determining LLMs' knowledge gaps, we diagnostically create unanswerable questions containing non-existent concepts or false premises, ensuring that they are outside the LLMs' vast training data. By compiling a benchmark, UnknownBench, which consists of both unanswerable and answerable questions, we quantitatively evaluate the LLMs' performance in maintaining honesty while being helpful. Using a model-agnostic unified confidence elicitation approach, we observe that most LLMs fail to consistently refuse or express uncertainty towards questions outside their parametric knowledge, although instruction fine-tuning and alignment techniques can provide marginal enhancements. Moreover, LLMs' uncertainty expression does not always stay consistent with the perceived confidence of their textual outputs.

Yiqing Xie, Emmy Liu, Gaokai Zhang et al.

When assessing the quality of coding agents, predominant benchmarks focus on solving single issues on GitHub, such as SWE-Bench. In contrast, in real use, these agents solve more various and complex tasks that involve other skills such as exploring codebases, testing software, and designing architecture. In this paper, we first characterize some transferable skills that are shared across diverse tasks by decomposing trajectories into fine-grained components, and derive a set of principles for designing auxiliary training tasks to teach language models these skills. Guided by these principles, we propose a training environment, Hybrid-Gym, consisting of a set of scalable synthetic tasks, such as function localization and dependency search. Experiments show that agents trained on our synthetic tasks effectively generalize to diverse real-world tasks that are not present in training, improving a base model by 25.4% absolute gain on SWE-Bench Verified, 7.9% on SWT-Bench Verified, and 5.1% on Commit-0 Lite. Hybrid-Gym also complements datasets built for the downstream tasks (e.g., improving SWE-Play by 4.9% on SWT-Bench Verified). Code available at: https://github.com/yiqingxyq/Hybrid-Gym.

Jiankai Tang, Tao Zhang, Jia Li et al.

Portable physiological monitoring is essential for early detection and management of cardiovascular disease, but current methods often require specialized equipment that limits accessibility or impose impractical postures that patients cannot maintain. Video-based photoplethysmography on smartphones offers a convenient noninvasive alternative, yet it still faces reliability challenges caused by motion artifacts, lighting variations, and single-view constraints. Few studies have demonstrated reliable application to cardiovascular patients, and no widely used open datasets exist for cross-device accuracy. To address these limitations, we introduce the M3PD dataset, the first publicly available dual-view mobile photoplethysmography dataset, comprising synchronized facial and fingertip videos captured simultaneously via front and rear smartphone cameras from 60 participants (including 47 cardiovascular patients). Building on this dual-view setting, we further propose F3Mamba, which fuses the facial and fingertip views through Mamba-based temporal modeling. The model reduces heart-rate error by 21.9 to 30.2 percent over existing single-view baselines while improving robustness in challenging real-world scenarios. Data and code: https://github.com/Health-HCI-Group/F3Mamba.

Xingyao Wang, Yangyi Chen, Lifan Yuan et al.

Large Language Model (LLM) agents, capable of performing a broad range of actions, such as invoking tools and controlling robots, show great potential in tackling real-world challenges. LLM agents are typically prompted to produce actions by generating JSON or text in a pre-defined format, which is usually limited by constrained action space (e.g., the scope of pre-defined tools) and restricted flexibility (e.g., inability to compose multiple tools). This work proposes to use executable Python code to consolidate LLM agents' actions into a unified action space (CodeAct). Integrated with a Python interpreter, CodeAct can execute code actions and dynamically revise prior actions or emit new actions upon new observations through multi-turn interactions. Our extensive analysis of 17 LLMs on API-Bank and a newly curated benchmark shows that CodeAct outperforms widely used alternatives (up to 20% higher success rate). The encouraging performance of CodeAct motivates us to build an open-source LLM agent that interacts with environments by executing interpretable code and collaborates with users using natural language. To this end, we collect an instruction-tuning dataset CodeActInstruct that consists of 7k multi-turn interactions using CodeAct. We show that it can be used with existing data to improve models in agent-oriented tasks without compromising their general capability. CodeActAgent, finetuned from Llama2 and Mistral, is integrated with Python interpreter and uniquely tailored to perform sophisticated tasks (e.g., model training) using existing libraries and autonomously self-debug.

Sha Li, Chi Han, Pengfei Yu et al.

The recent explosion of performance of large language models (LLMs) has changed the field of Natural Language Processing (NLP) more abruptly and seismically than any other shift in the field's 80-year history. This has resulted in concerns that the field will become homogenized and resource-intensive. The new status quo has put many academic researchers, especially PhD students, at a disadvantage. This paper aims to define a new NLP playground by proposing 20+ PhD-dissertation-worthy research directions, covering theoretical analysis, new and challenging problems, learning paradigms, and interdisciplinary applications.

Lifan Yuan, Ganqu Cui, Hanbin Wang et al. · tencent-ai, tsinghua

We introduce Eurus, a suite of large language models (LLMs) optimized for reasoning. Finetuned from Mistral-7B and CodeLlama-70B, Eurus models achieve state-of-the-art results among open-source models on a diverse set of benchmarks covering mathematics, code generation, and logical reasoning problems. Notably, Eurus-70B beats GPT-3.5 Turbo in reasoning through a comprehensive benchmarking across 12 tests covering five tasks, and achieves a 33.3% pass@1 accuracy on LeetCode and 32.6% on TheoremQA, two challenging benchmarks, substantially outperforming existing open-source models by margins more than 13.3%. The strong performance of Eurus can be primarily attributed to UltraInteract, our newly-curated large-scale, high-quality alignment dataset specifically designed for complex reasoning tasks. UltraInteract can be used in both supervised fine-tuning and preference learning. For each instruction, it includes a preference tree consisting of (1) reasoning chains with diverse planning strategies in a unified format, (2) multi-turn interaction trajectories with the environment and the critique, and (3) pairwise data to facilitate preference learning. UltraInteract allows us to conduct an in-depth exploration of preference learning for reasoning tasks. Our investigation reveals that some well-established preference learning algorithms may be less suitable for reasoning tasks compared to their effectiveness in general conversations. Inspired by this, we derive a novel reward modeling objective which, together with UltraInteract, leads to a strong reward model.

Alejandro Cuadron, Dacheng Li, Wenjie Ma et al.

Large Reasoning Models (LRMs) represent a breakthrough in AI problem-solving capabilities, but their effectiveness in interactive environments can be limited. This paper introduces and analyzes overthinking in LRMs. A phenomenon where models favor extended internal reasoning chains over environmental interaction. Through experiments on software engineering tasks using SWE Bench Verified, we observe three recurring patterns: Analysis Paralysis, Rogue Actions, and Premature Disengagement. We propose a framework to study these behaviors, which correlates with human expert assessments, and analyze 4018 trajectories. We observe that higher overthinking scores correlate with decreased performance, with reasoning models exhibiting stronger tendencies toward overthinking compared to non-reasoning models. Our analysis reveals that simple efforts to mitigate overthinking in agentic environments, such as selecting the solution with the lower overthinking score, can improve model performance by almost 30% while reducing computational costs by 43%. These results suggest that mitigating overthinking has strong practical implications. We suggest that by leveraging native function-calling capabilities and selective reinforcement learning overthinking tendencies could be mitigated. We also open-source our evaluation framework and dataset to facilitate research in this direction at https://github.com/AlexCuadron/Overthinking.

Weiwei Sun, Xuhui Zhou, Weihua Du et al.

While existing work focuses primarily on task success, we argue that effective real-world agents require optimizing three dimensions: productivity (task completion), proactivity (asking essential questions), and personalization (adapting to diverse user preferences). We introduce UserVille, an interactive environment with LLM-based user simulators enabling diverse, configurable user preferences. Leveraging UserVille, we introduce PPP, a multi-objective reinforcement learning approach that jointly optimizes all three dimensions: Productivity, Proactivity, and Personalization. Experiments on software engineering and deep research tasks show that agents trained with PPP achieve substantial improvements over strong baselines such as GPT-5 (+21.6 on average), demonstrating the ability to ask strategic clarifying questions, adapt to unseen user preferences, and improve task success through better interaction. This work demonstrates that explicitly optimizing for user-centered interaction is critical for building practical and effective AI agents.

Yue Pan, Xingyao Wang, Hanyue Zhang et al.

Remote monitoring of heart failure (HF) via speech signals provides a non-invasive and cost-effective solution for long-term patient management. However, substantial inter-individual heterogeneity in vocal characteristics often limits the accuracy of traditional cross-sectional classification models. To address this, we propose a Longitudinal Intra-Patient Tracking (LIPT) scheme designed to capture the trajectory of relative symptomatic changes within individuals. Central to this framework is a Personalised Sequential Encoder (PSE), which transforms longitudinal speech recordings into context-aware latent representations. By incorporating historical data at each timestamp, the PSE facilitates a holistic assessment of the clinical trajectory rather than modelling discrete visits independently. Experimental results from a cohort of 225 patients demonstrate that the LIPT paradigm significantly outperforms the classic cross-sectional approaches, achieving a recognition accuracy of 99.7% for clinical status transitions. The model's high sensitivity was further corroborated by additional follow-up data, confirming its efficacy in predicting HF deterioration and its potential to secure patient safety in remote, home-based settings. Furthermore, this work addresses the gap in existing literature by providing a comprehensive analysis of different speech task designs and acoustic features. Taken together, the superior performance of the LIPT framework and PSE architecture validates their readiness for integration into long-term telemonitoring systems, offering a scalable solution for remote heart failure management.

Gangda Deng, Zhaoling Chen, Zhongming Yu et al.

With AI agents increasingly deployed as long-running systems, it becomes essential to autonomously construct and continuously evolve customized software to enable interaction within dynamic environments. Yet, existing benchmarks evaluate agents on isolated, one-off coding tasks, neglecting the temporal dependencies and technical debt inherent in real-world software evolution. To bridge this gap, we introduce DeepCommit, an agentic pipeline that reconstructs verifiable Milestone DAGs from noisy commit logs, where milestones are defined as semantically cohesive development goals. These executable sequences enable EvoClaw, a novel benchmark that requires agents to sustain system integrity and limit error accumulation, dimensions of long-term software evolution largely missing from current benchmarks. Our evaluation of 12 frontier models across 4 agent frameworks reveals a critical vulnerability: overall performance scores drop significantly from $>$80% on isolated tasks to at most 38% in continuous settings, exposing agents' profound struggle with long-term maintenance and error propagation.

Zhaoling Chen, Xiangru Tang, Gangda Deng et al.

Code localization--identifying precisely where in a codebase changes need to be made--is a fundamental yet challenging task in software maintenance. Existing approaches struggle to efficiently navigate complex codebases when identifying relevant code sections. The challenge lies in bridging natural language problem descriptions with the appropriate code elements, often requiring reasoning across hierarchical structures and multiple dependencies. We introduce LocAgent, a framework that addresses code localization through graph-based representation. By parsing codebases into directed heterogeneous graphs, LocAgent creates a lightweight representation that captures code structures (files, classes, functions) and their dependencies (imports, invocations, inheritance), enabling LLM agents to effectively search and locate relevant entities through powerful multi-hop reasoning. Experimental results on real-world benchmarks demonstrate that our approach significantly enhances accuracy in code localization. Notably, our method with the fine-tuned Qwen-2.5-Coder-Instruct-32B model achieves comparable results to SOTA proprietary models at greatly reduced cost (approximately 86% reduction), reaching up to 92.7% accuracy on file-level localization while improving downstream GitHub issue resolution success rates by 12% for multiple attempts (Pass@10). Our code is available at https://github.com/gersteinlab/LocAgent.

Xingyao Wang, Simon Rosenberg, Juan Michelini et al.

Agents are now used widely in the process of software development, but building production-ready software engineering agents is a complex task. Deploying software agents effectively requires flexibility in implementation and experimentation, reliable and secure execution, and interfaces for users to interact with agents. In this paper, we present the OpenHands Software Agent SDK, a toolkit for implementing software development agents that satisfy these desiderata. This toolkit is a complete architectural redesign of the agent components of the popular OpenHands framework for software development agents, which has 64k+ GitHub stars. To achieve flexibility, we design a simple interface for implementing agents that requires only a few lines of code in the default case, but is easily extensible to more complex, full-featured agents with features such as custom tools, memory management, and more. For security and reliability, it delivers seamless local-to-remote execution portability, integrated REST/WebSocket services. For interaction with human users, it can connect directly to a variety of interfaces, such as visual workspaces (VS Code, VNC, browser), command-line interfaces, and APIs. Compared with existing SDKs from OpenAI, Claude, and Google, OpenHands uniquely integrates native sandboxed execution, lifecycle control, model-agnostic multi-LLM routing, and built-in security analysis. Empirical results on SWE-Bench Verified and GAIA benchmarks demonstrate strong performance. Put together, these elements allow the OpenHands Software Agent SDK to provide a practical foundation for prototyping, unlocking new classes of custom applications, and reliably deploying agents at scale.

Ke Yang, Jiateng Liu, John Wu et al.

The prominent large language models (LLMs) of today differ from past language models not only in size, but also in the fact that they are trained on a combination of natural language and formal language (code). As a medium between humans and computers, code translates high-level goals into executable steps, featuring standard syntax, logical consistency, abstraction, and modularity. In this survey, we present an overview of the various benefits of integrating code into LLMs' training data. Specifically, beyond enhancing LLMs in code generation, we observe that these unique properties of code help (i) unlock the reasoning ability of LLMs, enabling their applications to a range of more complex natural language tasks; (ii) steer LLMs to produce structured and precise intermediate steps, which can then be connected to external execution ends through function calls; and (iii) take advantage of code compilation and execution environment, which also provides diverse feedback for model improvement. In addition, we trace how these profound capabilities of LLMs, brought by code, have led to their emergence as intelligent agents (IAs) in situations where the ability to understand instructions, decompose goals, plan and execute actions, and refine from feedback are crucial to their success on downstream tasks. Finally, we present several key challenges and future directions of empowering LLMs with code.

Valerie Chen, Rohit Malhotra, Xingyao Wang et al. · cmu

LLM-powered agents are both a promising new technology and a source of complexity, where choices about models, tools, and prompting can affect their usefulness. While numerous benchmarks measure agent accuracy across domains, they mostly assume full automation, failing to represent the collaborative nature of real-world use cases. In this paper, we make two major steps towards the rigorous assessment of human-agent interactions. First, we propose PULSE, a framework for more efficient human-centric evaluation of agent designs, which comprises collecting user feedback, training an ML model to predict user satisfaction, and computing results by combining human satisfaction ratings with model-generated pseudo-labels. Second, we deploy the framework on a large-scale web platform built around the open-source software agent OpenHands, collecting in-the-wild usage data across over 15k users. We conduct case studies around how three agent design decisions -- choice of LLM backbone, planning strategy, and memory mechanisms -- impact developer satisfaction rates, yielding practical insights for software agent design. We also show how our framework can lead to more robust conclusions about agent design, reducing confidence intervals by 40% compared to a standard A/B test. Finally, we find substantial discrepancies between in-the-wild results and benchmark performance (e.g., the anti-correlation between results comparing claude-sonnet-4 and gpt-5), underscoring the limitations of benchmark-driven evaluation. Our findings provide guidance for evaluations of LLM agents with humans and identify opportunities for better agent designs.

Jiayi Pan, Xingyao Wang, Graham Neubig et al. · berkeley

We present SWE-Gym, the first environment for training real-world software engineering (SWE) agents. SWE-Gym contains 2,438 real-world Python task instances, each comprising a codebase with an executable runtime environment, unit tests, and a task specified in natural language. We use SWE-Gym to train language model based SWE agents, achieving up to 19% absolute gains in resolve rate on the popular SWE-Bench Verified and Lite test sets. We also experiment with inference-time scaling through verifiers trained on agent trajectories sampled from SWE-Gym. When combined with our fine-tuned SWE agents, we achieve 32.0% and 26.0% on SWE-Bench Verified and Lite, respectively, reflecting a new state-of-the-art for open-weight SWE agents. To facilitate further research, we publicly release SWE-Gym, models, and agent trajectories.

Abhinav Rastogi, Adam Yang, Albert Q. Jiang et al. · deepmind

We introduce Devstral-Small, a lightweight open source model for code agents with the best performance among models below 100B size. In this technical report, we give an overview of how we design and develop a model and craft specializations in agentic software development. The resulting model, Devstral-Small is a small 24B model, fast and easy to serve. Despite its size, Devstral-Small still attains competitive performance compared to models more than an order of magnitude larger.

Longju Bai, Zhemin Huang, Xingyao Wang et al.

The wide adoption of AI agents in complex human workflows is driving rapid growth in LLM token consumption. When agents are deployed on tasks that require a significant amount of tokens, three questions naturally arise: (1) Where do AI agents spend the tokens? (2) Which models are more token-efficient? and (3) Can agents predict their token usage before task execution? In this paper, we present the first systematic study of token consumption patterns in agentic coding tasks. We analyze trajectories from eight frontier LLMs on SWE-bench Verified and evaluate models' ability to predict their own token costs before task execution. We find that: (1) agentic tasks are uniquely expensive, consuming 1000x more tokens than code reasoning and code chat, with input tokens rather than output tokens driving the overall cost; (2) token usage is highly variable and inherently stochastic: runs on the same task can differ by up to 30x in total tokens, and higher token usage does not translate into higher accuracy; instead, accuracy often peaks at intermediate cost and saturates at higher costs; (3) models vary substantially in token efficiency: on the same tasks, Kimi-K2 and Claude-Sonnet-4.5, on average, consume over 1.5 million more tokens than GPT-5; (4) task difficulty rated by human experts only weakly aligns with actual token costs, revealing a fundamental gap between human-perceived complexity and the computational effort agents actually expend; and (5) frontier models fail to accurately predict their own token usage (with weak-to-moderate correlations, up to 0.39) and systematically underestimate real token costs. Our study offers new insights into the economics of AI agents and can inspire future research in this direction.

Xiangru Tang, Tianrui Qin, Tianhao Peng et al.

AI agent frameworks operate in isolation, forcing agents to rediscover solutions and repeat mistakes across different systems. Despite valuable problem-solving experiences accumulated by frameworks like smolagents, OpenHands, and OWL, this knowledge remains trapped within individual systems, preventing the emergence of collective intelligence. Current memory systems focus on individual agents or framework-specific demonstrations, failing to enable cross-architecture knowledge transfer. We introduce AGENT KB, a universal memory infrastructure enabling seamless experience sharing across heterogeneous agent frameworks without retraining. AGENT KB aggregates trajectories into a structured knowledge base and serves lightweight APIs. At inference time, hybrid retrieval operates through two stages: planning seeds agents with cross-domain workflows, while feedback applies targeted diagnostic fixes. A disagreement gate ensures retrieved knowledge enhances rather than disrupts reasoning, addressing knowledge interference in cross-framework transfer. We validate AGENT KB across major frameworks on GAIA, Humanity's Last Exam, GPQA, and SWE-bench. Results show substantial improvements across diverse model families: compared to baseline pass@1, smolagents with AGENT KB achieve up to 18.7pp gains at pass@3 (55.2% -> 73.9%), while OpenHands improves 4.0pp on SWE-bench pass@1 (24.3% -> 28.3%). Similar improvements are observed across all base model families. Ablations confirm that hybrid retrieval and feedback stages are essential, with automatically generated experiences matching manual curation. This establishes the foundation for collective agent intelligence through shared memory infrastructures.

Aditya Bharat Soni, Boxuan Li, Xingyao Wang et al. · cmu

Modern human labor is characterized by specialization; we train for years and develop particular tools that allow us to perform well across a variety of tasks. In addition, AI agents have been specialized for domains such as software engineering, web navigation, and workflow automation. However, this results in agents that are good for one thing but fail to generalize beyond their intended scope. One reason for this is that agent developers provide a highly specialized set of tools or make architectural decisions optimized for a specific use case or benchmark. In this work, we ask the question: what is the minimal set of general tools that can be used to achieve high performance across a diverse set of tasks? Our answer is OpenHands-Versa, a generalist agent built with a modest number of general tools: code editing and execution, web search, as well as multimodal web browsing and file access. Importantly, OpenHands-Versa demonstrates superior or competitive performance over leading specialized agents across three diverse and challenging benchmarks: SWE-Bench Multimodal, GAIA, and The Agent Company, outperforming the best-performing previously published results with absolute improvements in success rate of 9.1, 1.3, and 9.1 points respectively. Further, we show how existing state-of-the-art multi-agent systems fail to generalize beyond their target domains. These results demonstrate the feasibility of developing a generalist agent to solve diverse tasks and establish OpenHands-Versa as a strong baseline for future research.

Zhenhailong Wang, Joy Hsu, Xingyao Wang et al.

Despite significant advancements, large multimodal models (LMMs) still struggle to bridge the gap between low-level visual perception -- focusing on shapes, sizes, and layouts -- and high-level language reasoning, such as semantics and logic. This limitation is evident in tasks that require precise visual perception, like comparing geometric properties or solving visual reasoning problems. To study this failure mode, we focus on vector graphics -- images composed of 2D objects and shapes, prevalent in LMM-based tasks in web, design, and OS environments. We identify two key research questions: how can we enable precise visual perception, and how can we facilitate high-level reasoning based on such low-level perceptions? To capture fine visual details, we use Scalable Vector Graphics (SVG) for accurate encoding of visual scenes. However, SVGs are not readily interpretable by LMMs in a zero-shot manner. To tackle this, we propose the Visually Descriptive Language Model (VDLM), which introduces a Primal Visual Description (PVD) as an intermediate textual representation. PVD translates SVGs into a text-based abstraction consisting of primitive attributes (e.g., shape, position, measurement) and their corresponding values. PVD can be learned using task-agnostic synthesized data and represents visual primitives that are universal across vector graphics. This abstraction is more structured, allowing for direct interpretation by foundation models for zero-shot generalization. Without human-annotated data, empirical results show that VDLM significantly improves state-of-the-art LMMs like GPT-4o on various multimodal perception and reasoning tasks. Extensive analyses of VDLM show improved interpretability due to its disentangled perception and reasoning. We also demonstrate a positive correlation between PVD quality and task performance. Project page: https://mikewangwzhl.github.io/VDLM/

Xuehang Guo, Xingyao Wang, Yangyi Chen et al.

Software engineering (SE) is increasingly collaborative, with developers working together on shared complex codebases. Effective collaboration in shared environments requires participants -- whether humans or AI agents -- to stay on the same page as their environment evolves. When a collaborator's understanding diverges from the current state -- what we term the out-of-sync challenge -- the collaborator's actions may fail, leading to integration issues. In this work, we introduce SyncMind, a framework that systematically defines the out-of-sync problem faced by large language model (LLM) agents in collaborative software engineering (CSE). Based on SyncMind, we create SyncBench, a benchmark featuring 24,332 instances of agent out-of-sync scenarios in real-world CSE derived from 21 popular GitHub repositories with executable verification tests. Experiments on SyncBench uncover critical insights into existing LLM agents' capabilities and limitations. Besides substantial performance gaps among agents (from Llama-3.1 agent <= 3.33% to Claude-3.5-Sonnet >= 28.18%), their consistently low collaboration willingness (<= 4.86%) suggests fundamental limitations of existing LLM in CSE. However, when collaboration occurs, it positively correlates with out-of-sync recovery success. Minimal performance differences in agents' resource-aware out-of-sync recoveries further reveal their significant lack of resource awareness and adaptability, shedding light on future resource-efficient collaborative systems. Code and data are openly available on our project website: https://xhguo7.github.io/SyncMind/.

Junlong Li, Wenshuo Zhao, Jian Zhao et al. · cmu

Real-world language agents must handle complex, multi-step workflows across diverse Apps. For instance, an agent may manage emails by coordinating with calendars and file systems, or monitor a production database to detect anomalies and generate reports following an operating manual. However, existing language agent benchmarks often focus on narrow domains or simplified tasks that lack the diversity, realism, and long-horizon complexity required to evaluate agents' real-world performance. To address this gap, we introduce the Tool Decathlon (dubbed as Toolathlon), a benchmark for language agents offering diverse Apps and tools, realistic environment setup, and reliable execution-based evaluation. Toolathlon spans 32 software applications and 604 tools, ranging from everyday platforms such as Google Calendar and Notion to professional ones like WooCommerce, Kubernetes, and BigQuery. Most of the tools are based on a high-quality set of Model Context Protocol (MCP) servers that we may have revised or implemented ourselves. Unlike prior works, which primarily ensure functional realism but offer limited environment state diversity, we provide realistic initial environment states from real software, such as Canvas courses with dozens of students or real financial spreadsheets. This benchmark includes 108 manually sourced or crafted tasks in total, requiring interacting with multiple Apps over around 20 turns on average to complete. Each task is strictly verifiable through dedicated evaluation scripts. Comprehensive evaluation of SOTA models highlights their significant shortcomings: the best-performing model, Claude-4.5-Sonnet, achieves only a 38.6% success rate with 20.2 tool calling turns on average, while the top open-weights model DeepSeek-V3.2-Exp reaches 20.1%. We expect Toolathlon to drive the development of more capable language agents for real-world, long-horizon task execution.

Xuhui Zhou, Valerie Chen, Zora Zhiruo Wang et al. · cmu

Recent advances in coding agents have made them capable of planning, editing, running, and testing complex code bases. Despite their growing ability in coding tasks, these systems still struggle to infer and track user intent, especially when instructions are underspecified or context-dependent. To bridge this gap, we introduce ToM-SWE, a dual-agent architecture that pairs a primary software-engineering (SWE) agent with a lightweight theory-of-mind (ToM) partner agent dedicated to modeling the user's mental state. The ToM agent infers user goals, constraints, and preferences from instructions and interaction history, maintains a \textbf{persistent memory} of the user, and provides user-related suggestions to the SWE agent. In two software engineering benchmarks (ambiguous SWE-bench and stateful SWE-bench), ToM-SWE improves task success rates and user satisfaction. Notably, on the stateful SWE benchmark, a newly introduced evaluation that provides agents with a user simulator along with previous interaction histories, ToM-SWE achieves a substantially higher task success rate of 59.7\% compared to 18.1\% for OpenHands, a state-of-the-art SWE agent. Furthermore, in a three-week study with professional developers using ToM-SWE in their daily work, participants found it useful 86\% of the time, underscoring the value of stateful user modeling for practical coding agents.

Yueqi Song, Ketan Ramaneti, Zaid Sheikh et al.

Public research results on large-scale supervised finetuning of AI agents remain relatively rare, since the collection of agent training data presents unique challenges. In this work, we argue that the bottleneck is not a lack of underlying data sources, but that a large variety of data is fragmented across heterogeneous formats, tools, and interfaces. To this end, we introduce the agent data protocol (ADP), a light-weight representation language that serves as an "interlingua" between agent datasets in diverse formats and unified agent training pipelines downstream. The design of ADP is expressive enough to capture a large variety of tasks, including API/tool use, browsing, coding, software engineering, and general agentic workflows, while remaining simple to parse and train on without engineering at a per-dataset level. In experiments, we unified a broad collection of 13 existing agent training datasets into ADP format, and converted the standardized ADP data into training-ready formats for multiple agent frameworks. We performed SFT on these data, and demonstrated an average performance gain of ~20% over corresponding base models, and delivers state-of-the-art or near-SOTA performance on standard coding, browsing, tool use, and research benchmarks, without domain-specific tuning. All code and data are released publicly, in the hope that ADP could help lower the barrier to standardized, scalable, and reproducible agent training.

Rui Feng, Zhiyao Luo, Wei Wang et al.

Automatic assessment of cognitive impairment from spontaneous speech offers a promising, non-invasive avenue for early cognitive screening. However, current approaches often lack generalizability when deployed across different languages and clinical settings, limiting their practical utility. In this study, we propose CogBench, the first benchmark designed to evaluate the cross-lingual and cross-site generalizability of large language models (LLMs) for speech-based cognitive impairment assessment. Using a unified multimodal pipeline, we evaluate model performance on three speech datasets spanning English and Mandarin: ADReSSo, NCMMSC2021-AD, and a newly collected test set, CIR-E. Our results show that conventional deep learning models degrade substantially when transferred across domains. In contrast, LLMs equipped with chain-of-thought prompting demonstrate better adaptability, though their performance remains sensitive to prompt design. Furthermore, we explore lightweight fine-tuning of LLMs via Low-Rank Adaptation (LoRA), which significantly improves generalization in target domains. These findings offer a critical step toward building clinically useful and linguistically robust speech-based cognitive assessment tools.

Xingyao Wang, Hao Peng, Reyhaneh Jabbarvand et al.

Fine-tuning pre-trained language models (LMs) is essential for enhancing their capabilities. Existing techniques commonly fine-tune on input-output pairs (e.g., instruction tuning) or with numerical rewards that gauge the output quality (e.g., RLHF). We explore LMs' potential to learn from textual interactions (LETI) that not only check their correctness with binary labels but also pinpoint and explain errors in their outputs through textual feedback. Our focus is the code generation task, where the model produces code based on natural language instructions. This setting invites a natural and scalable way to acquire textual feedback: the error messages and stack traces from code execution using a Python interpreter. LETI iteratively fine-tunes the model, using the LM objective, on a concatenation of natural language instructions, LM-generated programs, and textual feedback. Prepended to this fine-tuning text, a binary reward token is used to differentiate correct and buggy solutions. LETI requires no ground-truth outputs for training and even outperforms a fine-tuned baseline that does. LETI not only improves the performance of LMs on a code generation dataset MBPP, but also generalizes to other datasets. Trained on MBPP, it achieves comparable or better performance than the base LMs on unseen problems in HumanEval. Furthermore, compared to binary feedback, we observe that textual feedback leads to improved generation quality and sample efficiency, achieving the same performance with fewer than half of the gradient steps. LETI is equally applicable in natural language tasks when they can be formulated as code generation, which we empirically verified on event argument extraction.

Xingyao Wang, David Jurgens

Online conversations include more than just text. Increasingly, image-based responses such as memes and animated gifs serve as culturally recognized and often humorous responses in conversation. However, while NLP has broadened to multimodal models, conversational dialog systems have largely focused only on generating text replies. Here, we introduce a new dataset of 1.56M text-gif conversation turns and introduce a new multimodal conversational model Pepe the King Prawn for selecting gif-based replies. We demonstrate that our model produces relevant and high-quality gif responses and, in a large randomized control trial of multiple models replying to real users, we show that our model replies with gifs that are significantly better received by the community.

Shaohuai Shi, Xianhao Zhou, Shutao Song et al.

Distributed training techniques have been widely deployed in large-scale deep neural networks (DNNs) training on dense-GPU clusters. However, on public cloud clusters, due to the moderate inter-connection bandwidth between instances, traditional state-of-the-art distributed training systems cannot scale well in training large-scale models. In this paper, we propose a new computing and communication efficient top-k sparsification communication library for distributed training. To further improve the system scalability, we optimize I/O by proposing a simple yet efficient multi-level data caching mechanism and optimize the update operation by introducing a novel parallel tensor operator. Experimental results on a 16-node Tencent Cloud cluster (each node with 8 Nvidia Tesla V100 GPUs) show that our system achieves 25%-40% faster than existing state-of-the-art systems on CNNs and Transformer. We finally break the record on DAWNBench on training ResNet-50 to 93% top-5 accuracy on ImageNet.

Gan Zhou, Zhi Li, Meng Fu et al.

With the increased demand on economy and efficiency of measurement technology, Non-Intrusive Load Monitoring (NILM) has received more and more attention as a cost-effective way to monitor electricity and provide feedback to users. Deep neural networks has been shown a great potential in the field of load disaggregation. In this paper, firstly, a new convolutional model based on residual blocks is proposed to avoid the degradation problem which traditional networks more or less suffer from when network layers are increased in order to learn more complex features. Secondly, we propose dilated convolution to curtail the excessive quantity of model parameters and obtain bigger receptive field, and multi-scale structure to learn mixed data features in a more targeted way. Thirdly, we give details about generating training and test set under certain rules. Finally, the algorithm is tested on real-house public dataset, UK-DALE, with three existing neural networks. The results are compared and analysed, the proposed model shows improvements on F1 score, MAE as well as model complexity across different appliances.

Xingyao Wang, Chengyu Liu, Yuwen Li et al.

Objective: This paper presents a novel heart sound segmentation algorithm based on Temporal-Framing Adaptive Network (TFAN), including state transition loss and dynamic inference for decoding the most likely state sequence. Methods: In contrast to previous state-of-the-art approaches, the TFAN-based method does not require any knowledge of the state duration of heart sounds and is therefore likely to generalize to non sinus rhythm. The TFAN-based method was trained on 50 recordings randomly chosen from Training set A of the 2016 PhysioNet/Computer in Cardiology Challenge and tested on the other 12 independent training and test databases (2099 recordings and 52180 beats). The databases for segmentation were separated into three levels of increasing difficulty (LEVEL-I, -II and -III) for performance reporting. Results: The TFAN-based method achieved a superior F1 score for all 12 databases except for `Test-B', with an average of 96.7%, compared to 94.6% for the state-of-the-art method. Moreover, the TFAN-based method achieved an overall F1 score of 99.2%, 94.4%, 91.4% on LEVEL-I, -II and -III data respectively, compared to 98.4%, 88.54% and 79.80% for the current state-of-the-art method. Conclusion: The TFAN-based method therefore provides a substantial improvement, particularly for more difficult cases, and on data sets not represented in the public training data. Significance: The proposed method is highly flexible and likely to apply to other non-stationary time series. Further work is required to understand to what extent this approach will provide improved diagnostic performance, although it is logical to assume superior segmentation will lead to improved diagnostics.