arXiv:2504.11436v4 Announce Type: replace-cross 
Abstract: We present evidence from a field experiment across 66 firms and 7,137 knowledge workers. Workers were randomly selected to access a generative AI tool integrated into applications they already used at work for email, meetings, and writing. In the second half of the 6-month experiment, the 80% of treated workers who used this tool spent two fewer hours on email each week and reduced their time working outside of regular hours. Apart from these individual time savings, we do not detect shifts in the quantity or composition of workers' tasks resulting from individual-level AI provision.

أظهر تجربة ميدانية حديثة شملت 66 شركة و7,137 عامل معرفة تأثير أداة الذكاء الاصطناعي التوليدي على أنماط العمل. وجدت الدراسة أن العمال الذين استخدموا أداة الذكاء الاصطناعي، المدمجة في التطبيقات التي يستخدمونها بالفعل في العمل للبريد الإلكتروني والاجتماعات والكتابة، وفروا في المتوسط ساعتين أسبوعياً في مهام البريد الإلكتروني. بالإضافة إلى ذلك، قلل هؤلاء العمال من وقت العمل خارج ساعات العمل العادية. ومع ذلك، لم تكشف التجربة عن أي تغييرات ملحوظة في كمية أو تكوين المهام التي يؤديها العمال.

Un reciente experimento de campo que involucró a 66 empresas y 7,137 trabajadores del conocimiento demostró el impacto de una herramienta de IA generativa en los patrones de trabajo. El estudio encontró que los trabajadores que utilizaron la herramienta de IA, integrada en sus aplicaciones existentes para correos electrónicos, reuniones y escritura, ahorraron un promedio de dos horas por semana en tareas de correo electrónico. Además, estos trabajadores redujeron su tiempo de trabajo fuera del horario regular. Sin embargo, el experimento no reveló cambios significativos en la cantidad o composición de las tareas realizadas por los trabajadores.

Une récente expérience de terrain impliquant 66 entreprises et 7 137 travailleurs du savoir a démontré l'impact d'un outil d'IA générative sur les modèles de travail. L'étude a révélé que les travailleurs utilisant l'outil d'IA, intégré dans leurs applications existantes pour les e-mails, les réunions et l'écriture, ont économisé en moyenne deux heures par semaine sur les tâches liées aux e-mails. De plus, ces travailleurs ont réduit leur temps de travail en dehors des heures normales. Cependant, l'expérience n'a pas révélé de changements significatifs dans la quantité ou la composition des tâches effectuées par les travailleurs.

A recent field experiment involving 66 firms and 7,137 knowledge workers demonstrated the impact of a generative AI tool on work patterns. The study found that workers who used the AI tool, integrated into their existing applications for email, meetings, and writing, saved an average of two hours per week on email tasks. Additionally, these workers reduced their time spent working outside regular hours. However, the experiment did not reveal any significant changes in the quantity or composition of tasks performed by the workers.

Shifting Work Patterns with Generative AI

arXiv:1906.07172v5 Announce Type: replace-cross 
Abstract: Equivariant neural networks are a class of neural networks designed to preserve symmetries inherent in the data. In this paper, we introduce a general method for modifying a neural network to enforce equivariance, a process we refer to as equivarification. We further show that group convolutional neural networks (G-CNNs) arise as a special case of our framework.

Equivariant neural networks and equivarification

arXiv:2311.13745v4 Announce Type: replace-cross 
Abstract: Diffusion models have become the most popular approach to deep generative modeling of images, largely due to their empirical performance and reliability. From a theoretical standpoint, a number of recent works have studied the iteration complexity of sampling, assuming access to an accurate diffusion model. In this work, we focus on understanding the sample complexity of training such a model; how many samples are needed to learn an accurate diffusion model using a sufficiently expressive neural network? Prior work showed bounds polynomial in the dimension, desired Total Variation error, and Wasserstein error. We show an exponential improvement in the dependence on Wasserstein error and depth, along with improved dependencies on other relevant parameters.

Improved Sample Complexity Bounds for Diffusion Model Training

arXiv:2410.22995v2 Announce Type: replace-cross 
Abstract: A hallmark of advanced artificial intelligence is the capacity to progress from passive visual perception to the strategic modification of visual information to facilitate complex reasoning. This advanced capability, however, remains critically underdeveloped in current Large Multi-modal Models (LMMs). The deficiency is often masked by evaluation metrics that prioritize final-answer accuracy, creating an illusion of competence where genuine reasoning is absent. Using the domain of geometric problem-solving as a precise instrument, we probe this issue through tasks that require constructing visual aids. To this end, we introduce \textbf{VisAidMath}, a challenging benchmark, and our novel Three-Layered Funnel Evaluation Framework. This framework moves beyond simple accuracy (ACCU) to scrutinize the generation of valid visual aids (PVA) and the soundness of subsequent reasoning steps (SPRS). Our extensive experiments on state-of-the-art models, including Doubao-Seed-1.6 and o4, reveal a profound ``Reasoning Illusion''. We observe that high surface-level accuracy conceals a catastrophic failure in the models' ability to produce valid visual aids or to reason from them. Our findings expose a fundamental schism between visual perception and logical deduction in modern LMMs. We host an evaluation platform at CodaBench for testing publicly. Homepage: https://nlp2ct.github.io/VisAidMathHomepage/ Evaluation: https://www.codabench.org/competitions/7634/

VisAidMath: Benchmarking Visual-Aided Mathematical Reasoning

arXiv:2508.01249v3 Announce Type: replace-cross 
Abstract: Large Language Model (LLM) agents offer a powerful new paradigm for solving various problems by combining natural language reasoning with the execution of external tools. However, their dynamic and non-transparent behavior introduces critical security risks, particularly in the presence of prompt injection attacks. In this work, we propose a novel insight that treats the agent runtime traces as structured programs with analyzable semantics. Thus, we present AgentArmor, a program analysis framework that converts agent traces into graph intermediate representation-based structured program dependency representations (e.g., CFG, DFG, and PDG) and enforces security policies via a type system. AgentArmor consists of three key components: (1) a graph constructor that reconstructs the agent's runtime traces as graph-based intermediate representations with control and data flow described within; (2) a property registry that attaches security-relevant metadata of interacted tools \& data, and (3) a type system that performs static inference and checking over the intermediate representation. By representing agent behavior as structured programs, AgentArmor enables program analysis for sensitive data flow, trust boundaries, and policy violations. We evaluate AgentArmor on the AgentDojo benchmark, the results show that AgentArmor can reduce the ASR to 3\%, with the utility drop only 1\%.

AgentArmor: Enforcing Program Analysis on Agent Runtime Trace to Defend Against Prompt Injection

arXiv:2511.14691v1 Announce Type: cross 
Abstract: Attention is the brain's ability to selectively focus on a few specific aspects while ignoring irrelevant ones. This biological principle inspired the attention mechanism in modern Transformers. Transformers now underpin large language models (LLMs) such as GPT, but at the cost of massive training and inference energy, leading to a large carbon footprint. While brain attention emerges from neural circuits, Transformer attention relies on dot-product similarity to weight elements in the input sequence. Neuromorphic computing, especially spiking neural networks (SNNs), offers a brain-inspired path to energy-efficient intelligence. Despite recent work on attention-based spiking Transformers, the core attention layer remains non-neuromorphic. Current spiking attention (i) relies on dot-product or element-wise similarity suited to floating-point operations, not event-driven spikes; (ii) keeps attention matrices that suffer from the von Neumann bottleneck, limiting in-memory computing; and (iii) still diverges from brain-like computation. To address these issues, we propose the Spiking STDP Transformer (S$^{2}$TDPT), a neuromorphic Transformer that implements self-attention through spike-timing-dependent plasticity (STDP), embedding query--key correlations in synaptic weights. STDP, a core mechanism of memory and learning in the brain and widely studied in neuromorphic devices, naturally enables in-memory computing and supports non-von Neumann hardware. On CIFAR-10 and CIFAR-100, our model achieves 94.35\% and 78.08\% accuracy with only four timesteps and 0.49 mJ on CIFAR-100, an 88.47\% energy reduction compared to a standard ANN Transformer. Grad-CAM shows that the model attends to semantically relevant regions, enhancing interpretability. Overall, S$^{2}$TDPT illustrates how biologically inspired attention can yield energy-efficient, hardware-friendly, and explainable neuromorphic models.

Attention via Synaptic Plasticity is All You Need: A Biologically Inspired Spiking Neuromorphic Transformer

arXiv:2511.08955v2 Announce Type: replace-cross 
Abstract: Simulating microstructure evolution (MicroEvo) is vital for materials design but demands high numerical accuracy, efficiency, and physical fidelity. Although recent studies on deep learning (DL) offer a promising alternative to traditional solvers, the field lacks standardized benchmarks. Existing studies are flawed due to a lack of comparing specialized MicroEvo DL models with state-of-the-art spatio-temporal architectures, an overemphasis on numerical accuracy over physical fidelity, and a failure to analyze error propagation over time. To address these gaps, we introduce MicroEvoEval, the first comprehensive benchmark for image-based microstructure evolution prediction. We evaluate 14 models, encompassing both domain-specific and general-purpose architectures, across four representative MicroEvo tasks with datasets specifically structured for both short- and long-term assessment. Our multi-faceted evaluation framework goes beyond numerical accuracy and computational cost, incorporating a curated set of structure-preserving metrics to assess physical fidelity. Our extensive evaluations yield several key insights. Notably, we find that modern architectures (e.g., VMamba), not only achieve superior long-term stability and physical fidelity but also operate with an order-of-magnitude greater computational efficiency. The results highlight the necessity of holistic evaluation and identify these modern architectures as a highly promising direction for developing efficient and reliable surrogate models in data-driven materials science.

MicroEvoEval: A Systematic Evaluation Framework for Image-Based Microstructure Evolution Prediction

arXiv:2508.18959v2 Announce Type: replace 
Abstract: Traditional map-making relies heavily on Geographic Information Systems (GIS), requiring domain expertise and being time-consuming, especially for repetitive tasks. Recent advances in generative AI (GenAI), particularly image diffusion models, offer new opportunities for automating and democratizing the map-making process. However, these models struggle with accurate map creation due to limited control over spatial composition and semantic layout. To address this, we integrate vector data to guide map generation in different styles, specified by the textual prompts. Our model is the first to generate accurate maps in controlled styles, and we have integrated it into a web application to improve its usability and accessibility. We conducted a user study with professional cartographers to assess the fidelity of generated maps, the usability of the web application, and the implications of ever-emerging GenAI in map-making. The findings have suggested the potential of our developed application and, more generally, the GenAI models in helping both non-expert users and professionals in creating maps more efficiently. We have also outlined further technical improvements and emphasized the new role of cartographers to advance the paradigm of AI-assisted map-making. The code and pre-trained models are available at https://github.com/claudaff/generative-ai-mapmaking/.

يتناول المقال دمج الذكاء الاصطناعي التوليدي في رسم الخرائط، مشيرًا إلى إمكانيته في أتمتة وتيسير عملية تعتمد تقليديًا على نظم المعلومات الجغرافية. على الرغم من التقدم، تواجه نماذج الذكاء الاصطناعي التوليدي تحديات في إنشاء خرائط دقيقة بسبب القيود في التركيب المكاني والتخطيط الدلالي. يقدم المؤلفون نموذجًا يولد خرائط دقيقة بأساليب محددة، تم التحقق منها من خلال دراسات مستخدمين مع خبراء الخرائط، مما يبرز آثار الذكاء الاصطناعي التوليدي في هذا المجال.

El artículo discute la integración de la IA generativa en la elaboración de mapas, destacando su potencial para automatizar y democratizar un proceso que tradicionalmente depende de los Sistemas de Información Geográfica (SIG). A pesar de los avances, los modelos de IA generativa enfrentan desafíos para crear mapas precisos debido a limitaciones en la composición espacial y el diseño semántico. Los autores presentan un modelo que genera mapas precisos en estilos controlados, validado a través de estudios de usuarios con cartógrafos profesionales, enfatizando las implicaciones de la IA generati…

L'article aborde l'intégration de l'IA générative dans la cartographie, soulignant son potentiel à automatiser et démocratiser un processus traditionnellement dépendant des systèmes d'information géographique (SIG). Malgré les avancées, les modèles d'IA générative rencontrent des défis pour créer des cartes précises en raison de limitations dans la composition spatiale et la mise en page sémantique. Les auteurs présentent un modèle qui génère des cartes précises dans des styles contrôlés, validé par des études utilisateurs avec des cartographes professionnels, soulignant les implications de l'…

The article discusses the integration of generative AI in map-making, highlighting its potential to automate and democratize the process traditionally reliant on Geographic Information Systems (GIS). Despite advancements, generative AI models face challenges in creating accurate maps due to limitations in spatial composition and semantic layout. The authors present a model that generates precise maps in controlled styles, validated through user studies with professional cartographers, emphasizing the implications of generative AI in the field.

Generative AI in Map-Making: A Technical Exploration and Its Implications for Cartographers

arXiv:2511.11208v1 Announce Type: new 
Abstract: Federated Learning (FL) enables collaborative model training across decentralized devices while preserving data privacy. However, FL methods typically run for a predefined number of global rounds, often leading to unnecessary computation when optimal performance is reached earlier. In addition, training may continue even when the model fails to achieve meaningful performance. To address this inefficiency, we introduce a zero-shot synthetic validation framework that leverages generative AI to monitor model performance and determine early stopping points. Our approach adaptively stops training near the optimal round, thereby conserving computational resources and enabling rapid hyperparameter adjustments. Numerical results on multi-label chest X-ray classification demonstrate that our method reduces training rounds by up to 74% while maintaining accuracy within 1% of the optimal.

يتيح التعلم الفيدرالي (FL) تدريب النماذج بشكل تعاوني عبر أجهزة موزعة مع الحفاظ على خصوصية البيانات. ومع ذلك، فإن طرق FL عادةً ما تعمل لعدد محدد مسبقًا من الجولات العالمية، مما يؤدي غالبًا إلى حسابات غير ضرورية عندما يتم الوصول إلى الأداء الأمثل في وقت مبكر. لمعالجة هذه الكفاءة، تم تقديم إطار عمل للتحقق الاصطناعي بدون طلقات يستخدم الذكاء الاصطناعي التوليدي لمراقبة أداء النموذج وتحديد نقاط التوقف المبكر، مما قد يقلل من جولات التدريب بنسبة تصل إلى 74% مع الحفاظ على الدقة ضمن 1% من الأمثل.

El Aprendizaje Federado (FL) permite el entrenamiento colaborativo de modelos a través de dispositivos descentralizados, garantizando la privacidad de los datos. Sin embargo, los métodos FL suelen ejecutarse durante un número predefinido de rondas globales, lo que puede llevar a cálculos innecesarios cuando se alcanza un rendimiento óptimo antes. Para abordar esta ineficiencia, se ha introducido un nuevo marco de validación sintética de cero disparos que utiliza IA generativa para monitorear el rendimiento del modelo y determinar puntos de parada temprana, lo que podría reducir las rondas de e…

L'apprentissage fédéré (FL) permet un entraînement collaboratif des modèles sur des dispositifs décentralisés tout en garantissant la confidentialité des données. Les méthodes FL traditionnelles fonctionnent souvent pendant un nombre prédéfini de tours globaux, ce qui peut entraîner des calculs inutiles lorsque des performances optimales sont atteintes plus tôt. Pour améliorer l'efficacité, un nouveau cadre de validation synthétique zéro-shot utilisant l'IA générative a été introduit pour surveiller les performances du modèle et déterminer les points d'arrêt précoces, réduisant potentiellement…

Federated Learning (FL) allows collaborative model training across decentralized devices while ensuring data privacy. Traditional FL methods often run for a set number of global rounds, which can lead to unnecessary computations when optimal performance is achieved earlier. To improve efficiency, a new zero-shot synthetic validation framework using generative AI has been introduced to monitor model performance and determine early stopping points, potentially reducing training rounds by up to 74% while maintaining accuracy within 1% of the optimal.

Shifting Work Patterns with Generative AI

Was this article worth reading? Share it