In recent years, two parallel research trends have emerged in machine learning, yet their intersections remain largely unexplored. On one hand, there has been a significant increase in literature focused on Individual Treatment Effect (ITE) modeling, particularly targeting the Conditional Average Treatment Effect (CATE) using meta-learner techniques. These approaches often aim to identify causal effects from observational data. On the other hand, the field of Explainable Machine Learning (XML) has gained traction, with various approaches developed to explain complex models and make their predictions more interpretable. A prominent technique in this area is Shapley Additive Explanations (SHAP), which has become mainstream in data science for analyzing supervised learning models. However, there has been limited exploration of SHAP application in identifying predictive biomarkers through CATE models, a crucial aspect in pharmaceutical precision medicine. We address inherent challenges associated with the SHAP concept in multi-stage CATE strategies and introduce a surrogate estimation approach that is agnostic to the choice of CATE strategy, effectively reducing computational burdens in high-dimensional data. Using this approach, we conduct simulation benchmarking to evaluate the ability to accurately identify biomarkers using SHAP values derived from various CATE meta-learners and Causal Forest.
The application of Large Language Models (LLMs) for Automated Algorithm Discovery (AAD), particularly for optimisation heuristics, is an emerging field of research. This emergence necessitates robust, standardised benchmarking practices to rigorously evaluate the capabilities and limitations of LLM-driven AAD methods and the resulting generated algorithms, especially given the opacity of their design process and known issues with existing benchmarks. To address this need, we introduce BLADE (Benchmark suite for LLM-driven Automated Design and Evolution), a modular and extensible framework specifically designed for benchmarking LLM-driven AAD methods in a continuous black-box optimisation context. BLADE integrates collections of benchmark problems (including MA-BBOB and SBOX-COST among others) with instance generators and textual descriptions aimed at capability-focused testing, such as generalisation, specialisation and information exploitation. It offers flexible experimental setup options, standardised logging for reproducibility and fair comparison, incorporates methods for analysing the AAD process (e.g., Code Evolution Graphs and various visualisation approaches) and facilitates comparison against human-designed baselines through integration with established tools like IOHanalyser and IOHexplainer. BLADE provides an `out-of-the-box' solution to systematically evaluate LLM-driven AAD approaches. The framework is demonstrated through two distinct use cases exploring mutation prompt strategies and function specialisation.
Pre-trained deep learning models, known as foundation models, have become essential building blocks in machine learning domains such as natural language processing and image domains. This trend has extended to respiratory and heart sound models, which have demonstrated effectiveness as off-the-shelf feature extractors. However, their evaluation benchmarking has been limited, resulting in incompatibility with state-of-the-art (SOTA) performance, thus hindering proof of their effectiveness. This study investigates the practical effectiveness of off-the-shelf audio foundation models by comparing their performance across four respiratory and heart sound tasks with SOTA fine-tuning results. Experiments show that models struggled on two tasks with noisy data but achieved SOTA performance on the other tasks with clean data. Moreover, general-purpose audio models outperformed a respiratory sound model, highlighting their broader applicability. With gained insights and the released code, we contribute to future research on developing and leveraging foundation models for respiratory and heart sounds.
Measuring progress is fundamental to the advancement of any scientific field. As benchmarks play an increasingly central role, they also grow more susceptible to distortion. Chatbot Arena has emerged as the go-to leaderboard for ranking the most capable AI systems. Yet, in this work we identify systematic issues that have resulted in a distorted playing field. We find that undisclosed private testing practices benefit a handful of providers who are able to test multiple variants before public release and retract scores if desired. We establish that the ability of these providers to choose the best score leads to biased Arena scores due to selective disclosure of performance results. At an extreme, we identify 27 private LLM variants tested by Meta in the lead-up to the Llama-4 release. We also establish that proprietary closed models are sampled at higher rates (number of battles) and have fewer models removed from the arena than open-weight and open-source alternatives. Both these policies lead to large data access asymmetries over time. Providers like Google and OpenAI have received an estimated 19.2% and 20.4% of all data on the arena, respectively. In contrast, a combined 83 open-weight models have only received an estimated 29.7% of the total data. We show that access to Chatbot Arena data yields substantial benefits; even limited additional data can result in relative performance gains of up to 112% on the arena distribution, based on our conservative estimates. Together, these dynamics result in overfitting to Arena-specific dynamics rather than general model quality. The Arena builds on the substantial efforts of both the organizers and an open community that maintains this valuable evaluation platform. We offer actionable recommendations to reform the Chatbot Arena's evaluation framework and promote fairer, more transparent benchmarking for the field
This work presents an RGB-D imaging-based approach to marker-free hand-eye calibration using a novel implementation of the iterative closest point (ICP) algorithm with a robust point-to-plane (PTP) objective formulated on a Lie algebra. Its applicability is demonstrated through comprehensive experiments using three well known serial manipulators and two RGB-D cameras. With only three randomly chosen robot configurations, our approach achieves approximately 90% successful calibrations, demonstrating 2-3x higher convergence rates to the global optimum compared to both marker-based and marker-free baselines. We also report 2 orders of magnitude faster convergence time (0.8 +/- 0.4 s) for 9 robot configurations over other marker-free methods. Our method exhibits significantly improved accuracy (5 mm in task space) over classical approaches (7 mm in task space) whilst being marker-free. The benchmarking dataset and code are open sourced under Apache 2.0 License, and a ROS 2 integration with robot abstraction is provided to facilitate deployment.
Perception systems in robotics encounter significant challenges in high-speed and dynamic conditions when relying on traditional cameras, where motion blur can compromise spatial feature integrity and task performance. Brain-inspired vision sensors (BVS) have recently gained attention as an alternative, offering high temporal resolution with reduced bandwidth and power requirements. Here, we present the first quantitative evaluation framework for two representative classes of BVSs in variable-speed robotic sensing, including event-based vision sensors (EVS) that detect asynchronous temporal contrasts, and the primitive-based sensor Tianmouc that employs a complementary mechanism to encode both spatiotemporal changes and intensity. A unified testing protocol is established, including crosssensor calibrations, standardized testing platforms, and quality metrics to address differences in data modality. From an imaging standpoint, we evaluate the effects of sensor non-idealities, such as motion-induced distortion, on the capture of structural information. For functional benchmarking, we examine task performance in corner detection and motion estimation under different rotational speeds. Results indicate that EVS performs well in highspeed, sparse scenarios and in modestly fast, complex scenes, but exhibits performance limitations in high-speed, cluttered settings due to pixel-level bandwidth variations and event rate saturation. In comparison, Tianmouc demonstrates consistent performance across sparse and complex scenarios at various speeds, supported by its global, precise, high-speed spatiotemporal gradient samplings. These findings offer valuable insights into the applicationdependent suitability of BVS technologies and support further advancement in this area.
Retrieval Augmented Generation (RAG) has become one of the most popular methods for bringing knowledge-intensive context to large language models (LLM) because of its ability to bring local context at inference time without the cost or data leakage risks associated with fine-tuning. A clear separation of private information from the LLM training has made RAG the basis for many enterprise LLM workloads as it allows the company to augment LLM's understanding using customers' private documents. Despite its popularity for private documents in enterprise deployments, current RAG benchmarks for validating and optimizing RAG pipelines draw their corpora from public data such as Wikipedia or generic web pages and offer little to no personal context. Seeking to empower more personal and private RAG we release the EnronQA benchmark, a dataset of 103,638 emails with 528,304 question-answer pairs across 150 different user inboxes. EnronQA enables better benchmarking of RAG pipelines over private data and allows for experimentation on the introduction of personalized retrieval settings over realistic data. Finally, we use EnronQA to explore the tradeoff in memorization and retrieval when reasoning over private documents.
Arabic text diacritization remains a persistent challenge in natural language processing due to the language's morphological richness. In this paper, we introduce Sadeed, a novel approach based on a fine-tuned decoder-only language model adapted from Kuwain 1.5B Hennara et al. [2025], a compact model originally trained on diverse Arabic corpora. Sadeed is fine-tuned on carefully curated, high-quality diacritized datasets, constructed through a rigorous data-cleaning and normalization pipeline. Despite utilizing modest computational resources, Sadeed achieves competitive results compared to proprietary large language models and outperforms traditional models trained on similar domains. Additionally, we highlight key limitations in current benchmarking practices for Arabic diacritization. To address these issues, we introduce SadeedDiac-25, a new benchmark designed to enable fairer and more comprehensive evaluation across diverse text genres and complexity levels. Together, Sadeed and SadeedDiac-25 provide a robust foundation for advancing Arabic NLP applications, including machine translation, text-to-speech, and language learning tools.
As large language models (LLMs) evolve into tool-using agents, the ability to browse the web in real-time has become a critical yardstick for measuring their reasoning and retrieval competence. Existing benchmarks such as BrowseComp concentrate on English and overlook the linguistic, infrastructural, and censorship-related complexities of other major information ecosystems -- most notably Chinese. To address this gap, we introduce BrowseComp-ZH, a high-difficulty benchmark purpose-built to comprehensively evaluate LLM agents on the Chinese web. BrowseComp-ZH consists of 289 multi-hop questions spanning 11 diverse domains. Each question is reverse-engineered from a short, objective, and easily verifiable answer (e.g., a date, number, or proper noun). A two-stage quality control protocol is applied to strive for high question difficulty and answer uniqueness. We benchmark over 20 state-of-the-art language models and agentic search systems on our proposed BrowseComp-ZH. Despite their strong conversational and retrieval capabilities, most models struggle severely: a large number achieve accuracy rates below 10%, and only a handful exceed 20%. Even the best-performing system, OpenAI's DeepResearch, reaches just 42.9%. These results demonstrate the considerable difficulty of BrowseComp-ZH, where success demands not only effective retrieval strategies, but also sophisticated reasoning and information reconciliation -- capabilities that current models still struggle to master. Our dataset, construction guidelines, and benchmark results have been publicly released at https://github.com/PALIN2018/BrowseComp-ZH.
Determining the most effective Large Language Model for code smell detection presents a complex challenge. This study introduces a structured methodology and evaluation matrix to tackle this issue, leveraging a curated dataset of code samples consistently annotated with known smells. The dataset spans four prominent programming languages Java, Python, JavaScript, and C++; allowing for cross language comparison. We benchmark two state of the art LLMs, OpenAI GPT 4.0 and DeepSeek-V3, using precision, recall, and F1 score as evaluation metrics. Our analysis covers three levels of detail: overall performance, category level performance, and individual code smell type performance. Additionally, we explore cost effectiveness by comparing the token based detection approach of GPT 4.0 with the pattern-matching techniques employed by DeepSeek V3. The study also includes a cost analysis relative to traditional static analysis tools such as SonarQube. The findings offer valuable guidance for practitioners in selecting an efficient, cost effective solution for automated code smell detection
As Natural Language Processing (NLP) models continue to evolve and become integral to high-stakes applications, ensuring their interpretability remains a critical challenge. Given the growing variety of explainability methods and diverse stakeholder requirements, frameworks that help stakeholders select appropriate explanations tailored to their specific use cases are increasingly important. To address this need, we introduce EvalxNLP, a Python framework for benchmarking state-of-the-art feature attribution methods for transformer-based NLP models. EvalxNLP integrates eight widely recognized explainability techniques from the Explainable AI (XAI) literature, enabling users to generate and evaluate explanations based on key properties such as faithfulness, plausibility, and complexity. Our framework also provides interactive, LLM-based textual explanations, facilitating user understanding of the generated explanations and evaluation outcomes. Human evaluation results indicate high user satisfaction with EvalxNLP, suggesting it is a promising framework for benchmarking explanation methods across diverse user groups. By offering a user-friendly and extensible platform, EvalxNLP aims at democratizing explainability tools and supporting the systematic comparison and advancement of XAI techniques in NLP.
We introduce OSVBench, a new benchmark for evaluating Large Language Models (LLMs) in generating complete specification code pertaining to operating system kernel verification tasks. The benchmark first defines the specification generation problem into a program synthesis problem within a confined scope of syntax and semantics by providing LLMs with the programming model. The LLMs are required to understand the provided verification assumption and the potential syntax and semantics space to search for, then generate the complete specification for the potentially buggy operating system code implementation under the guidance of the high-level functional description of the operating system. This benchmark is built upon a real-world operating system kernel, Hyperkernel, and consists of 245 complex specification generation tasks in total, each is a long context task of about 20k-30k tokens. Our comprehensive evaluation of 12 LLMs exhibits the limited performance of the current LLMs on the specification generation tasks for operating system verification. Significant disparities in their performance on the benchmark highlight differences in their ability to handle long-context code generation tasks. The evaluation toolkit and benchmark are available at https://github.com/lishangyu-hkust/OSVBench.
Nanobodies, single-domain antibody fragments derived from camelid heavy-chain-only antibodies, exhibit unique advantages such as compact size, high stability, and strong binding affinity, making them valuable tools in therapeutics and diagnostics. While recent advances in pretrained protein and antibody language models (PPLMs and PALMs) have greatly enhanced biomolecular understanding, nanobody-specific modeling remains underexplored and lacks a unified benchmark. To address this gap, we introduce NbBench, the first comprehensive benchmark suite for nanobody representation learning. Spanning eight biologically meaningful tasks across nine curated datasets, NbBench encompasses structure annotation, binding prediction, and developability assessment. We systematically evaluate eleven representative models--including general-purpose protein LMs, antibody-specific LMs, and nanobody-specific LMs--in a frozen setting. Our analysis reveals that antibody language models excel in antigen-related tasks, while performance on regression tasks such as thermostability and affinity remains challenging across all models. Notably, no single model consistently outperforms others across all tasks. By standardizing datasets, task definitions, and evaluation protocols, NbBench offers a reproducible foundation for assessing and advancing nanobody modeling.
Recent progress in Meta-Black-Box-Optimization (MetaBBO) has demonstrated that using RL to learn a meta-level policy for dynamic algorithm configuration (DAC) over an optimization task distribution could significantly enhance the performance of the low-level BBO algorithm. However, the online learning paradigms in existing works makes the efficiency of MetaBBO problematic. To address this, we propose an offline learning-based MetaBBO framework in this paper, termed Q-Mamba, to attain both effectiveness and efficiency in MetaBBO. Specifically, we first transform DAC task into long-sequence decision process. This allows us further introduce an effective Q-function decomposition mechanism to reduce the learning difficulty within the intricate algorithm configuration space. Under this setting, we propose three novel designs to meta-learn DAC policy from offline data: we first propose a novel collection strategy for constructing offline DAC experiences dataset with balanced exploration and exploitation. We then establish a decomposition-based Q-loss that incorporates conservative Q-learning to promote stable offline learning from the offline dataset. To further improve the offline learning efficiency, we equip our work with a Mamba architecture which helps long-sequence learning effectiveness and efficiency by selective state model and hardware-aware parallel scan respectively. Through extensive benchmarking, we observe that Q-Mamba achieves competitive or even superior performance to prior online/offline baselines, while significantly improving the training efficiency of existing online baselines. We provide sourcecodes of Q-Mamba at https://github.com/MetaEvo/Q-Mamba.
AI-generated synthetic media are increasingly used in real-world scenarios, often with the purpose of spreading misinformation and propaganda through social media platforms, where compression and other processing can degrade fake detection cues. Currently, many forensic tools fail to account for these in-the-wild challenges. In this work, we introduce TrueFake, a large-scale benchmarking dataset of 600,000 images including top notch generative techniques and sharing via three different social networks. This dataset allows for rigorous evaluation of state-of-the-art fake image detectors under very realistic and challenging conditions. Through extensive experimentation, we analyze how social media sharing impacts detection performance, and identify current most effective detection and training strategies. Our findings highlight the need for evaluating forensic models in conditions that mirror real-world use.
Synthetic image source attribution is an open challenge, with an increasing number of image generators being released yearly. The complexity and the sheer number of available generative techniques, as well as the scarcity of high-quality open source datasets of diverse nature for this task, make training and benchmarking synthetic image source attribution models very challenging. WILD is a new in-the-Wild Image Linkage Dataset designed to provide a powerful training and benchmarking tool for synthetic image attribution models. The dataset is built out of a closed set of 10 popular commercial generators, which constitutes the training base of attribution models, and an open set of 10 additional generators, simulating a real-world in-the-wild scenario. Each generator is represented by 1,000 images, for a total of 10,000 images in the closed set and 10,000 images in the open set. Half of the images are post-processed with a wide range of operators. WILD allows benchmarking attribution models in a wide range of tasks, including closed and open set identification and verification, and robust attribution with respect to post-processing and adversarial attacks. Models trained on WILD are expected to benefit from the challenging scenario represented by the dataset itself. Moreover, an assessment of seven baseline methodologies on closed and open set attribution is presented, including robustness tests with respect to post-processing.
Tabular data analysis presents unique challenges due to its heterogeneous feature types, missing values, and complex interactions. While traditional machine learning methods, such as gradient boosting, often outperform deep learning approaches, recent advancements in neural architectures offer promising alternatives. This paper introduces TabKAN, a novel framework that advances tabular data modeling using Kolmogorov-Arnold Networks (KANs). Unlike conventional deep learning models, KANs leverge learnable activation functions on edges, which improve both interpretability and training efficiency. Our contributions include: (1) the introduction of modular KAN-based architectures for tabular data analysis, (2) the development of a transfer learning framework for KAN models that supports knowledge transfer between domains, (3) the development of model-specific interpretability for tabular data learning, which reduces dependence on post hoc and model-agnostic analysis, and (4) comprehensive evaluation of vanilla supervised learning across binary and multi-class classification tasks. Through extensive benchmarking on diverse public datasets, TabKAN demonstrates superior performance in supervised learning while significantly outperforming classical and Transformer-based models in transfer learning scenarios. Our findings highlight the advantage of KAN-based architectures in transferring knowledge across domains and narrowing the gap between traditional machine learning and deep learning for structured data.
Tracking internal layers in radar echograms with high accuracy is essential for understanding ice sheet dynamics and quantifying the impact of accelerated ice discharge in Greenland and other polar regions due to contemporary global climate warming. Deep learning algorithms have become the leading approach for automating this task, but the absence of a standardized and well-annotated echogram dataset has hindered the ability to test and compare algorithms reliably, limiting the advancement of state-of-the-art methods for the radar echogram layer tracking problem. This study introduces the first comprehensive ``deep learning ready'' radar echogram dataset derived from Snow Radar airborne data collected during the National Aeronautics and Space Administration Operation Ice Bridge (OIB) mission in 2012. The dataset contains 13,717 labeled and 57,815 weakly-labeled echograms covering diverse snow zones (dry, ablation, wet) with varying along-track resolutions. To demonstrate its utility, we evaluated the performance of five deep learning models on the dataset. Our results show that while current computer vision segmentation algorithms can identify and track snow layer pixels in echogram images, advanced end-to-end models are needed to directly extract snow depth and annual accumulation from echograms, reducing or eliminating post-processing. The dataset and accompanying benchmarking framework provide a valuable resource for advancing radar echogram layer tracking and snow accumulation estimation, advancing our understanding of polar ice sheets response to climate warming.
Multimodal LLMs are turning their focus to video benchmarks, however most video benchmarks only provide outcome supervision, with no intermediate or interpretable reasoning steps. This makes it challenging to assess if models are truly able to combine perceptual and temporal information to reason about videos, or simply get the correct answer by chance or by exploiting linguistic biases. To remedy this, we provide a new video reasoning dataset called MINERVA for modern multimodal models. Each question in the dataset comes with 5 answer choices, as well as detailed, hand-crafted reasoning traces. Our dataset is multimodal, diverse in terms of video domain and length, and consists of complex multi-step questions. Extensive benchmarking shows that our dataset provides a challenge for frontier open-source and proprietary models. We perform fine-grained error analysis to identify common failure modes across various models, and create a taxonomy of reasoning errors. We use this to explore both human and LLM-as-a-judge methods for scoring video reasoning traces, and find that failure modes are primarily related to temporal localization, followed by visual perception errors, as opposed to logical or completeness errors. The dataset, along with questions, answer candidates and reasoning traces will be publicly available under https://github.com/google-deepmind/neptune?tab=readme-ov-file\#minerva.
Developing accurate and reliable models for forest types mapping is critical to support efforts for halting deforestation and for biodiversity conservation (such as European Union Deforestation Regulation (EUDR)). This work introduces ForTy, a benchmark for global-scale FORest TYpes mapping using multi-temporal satellite data1. The benchmark comprises 200,000 time series of image patches, each consisting of Sentinel-2, Sentinel-1, climate, and elevation data. Each time series captures variations at monthly or seasonal cadence. Per-pixel annotations, including forest types and other land use classes, support image segmentation tasks. Unlike most existing land use products that often categorize all forest areas into a single class, our benchmark differentiates between three forest types classes: natural forest, planted forest, and tree crops. By leveraging multiple public data sources, we achieve global coverage with this benchmark. We evaluate the forest types dataset using several baseline models, including convolution neural networks and transformer-based models. Additionally, we propose a novel transformer-based model specifically designed to handle multi-modal, multi-temporal satellite data for forest types mapping. Our experimental results demonstrate that the proposed model surpasses the baseline models in performance.