The concept of the Web of Agents (WoA), which transforms the static, document-centric Web into an environment of autonomous agents acting on users' behalf, has attracted growing interest as large language models (LLMs) become more capable. However, research in this area is still fragmented across different communities. Contemporary surveys catalog the latest LLM-powered frameworks, while the rich histories of Multi-Agent Systems (MAS) and the Semantic Web are often treated as separate, legacy domains. This fragmentation obscures the intellectual lineage of modern systems and hinders a holistic understanding of the field's trajectory. We present the first comprehensive evolutionary overview of the WoA. We show that modern protocols like A2A and the MCP, are direct evolutionary responses to the well-documented limitations of earlier standards like FIPA standards and OWL-based semantic agents. To systematize this analysis, we introduce a four-axis taxonomy (semantic foundation, communication paradigm, locus of intelligence, discovery mechanism). This framework provides a unified analytical lens for comparing agent architectures across all generations, revealing a clear line of descent where others have seen a disconnect. Our analysis identifies a paradigm shift in the 'locus of intelligence': from being encoded in external data (Semantic Web) or the platform (MAS) to being embedded within the agent's core model (LLM). This shift is foundational to modern Agentic AI, enabling the scalable and adaptive systems the WoA has long envisioned. We conclude that while new protocols are essential, they are insufficient for building a robust, open, trustworthy ecosystem. Finally, we argue that the next research frontier lies in solving persistent socio-technical challenges, and we map out a new agenda focused on decentralized identity, economic models, security, and governance for the emerging WoA.

The environmental impact of Artificial Intelligence (AI)-enabled systems is increasing rapidly, and software engineering plays a critical role in developing sustainable solutions. The "Greening AI with Software Engineering" CECAM-Lorentz workshop (no. 1358, 2025) funded by the Centre Européen de Calcul Atomique et Moléculaire and the Lorentz Center, provided an interdisciplinary forum for 29 participants, from practitioners to academics, to share knowledge, ideas, practices, and current results dedicated to advancing green software and AI research. The workshop was held February 3-7, 2025, in Lausanne, Switzerland. Through keynotes, flash talks, and collaborative discussions, participants identified and prioritized key challenges for the field. These included energy assessment and standardization, benchmarking practices, sustainability-aware architectures, runtime adaptation, empirical methodologies, and education. This report presents a research agenda emerging from the workshop, outlining open research directions and practical recommendations to guide the development of environmentally sustainable AI-enabled systems rooted in software engineering principles.

Time plays a critical role in how information is generated, retrieved, and interpreted. In this survey, we provide a comprehensive overview of Temporal Question Answering (TQA), a research area that focuses on answering questions involving temporal constraints or context. As the amount of time-stamped content from sources like news articles, web archives, and knowledge bases increases, systems must address challenges such as detecting temporal intent, normalizing time expressions, ordering events, and reasoning over evolving or ambiguous facts. We focus on recent advances in TQA enabled by neural architectures, especially transformer-based models and Large Language Models (LLMs), highlighting progress in temporal language modeling, retrieval-augmented generation (RAG), and temporal reasoning. We also discuss benchmark datasets and evaluation strategies designed to test temporal robustness, recency awareness, and generalization.

This paper quantitatively investigates the crash severity of Autonomous Vehicles (AVs) with spatially localized machine learning and macroscopic measures of the urban built environment. Extending beyond the microscopic effects of individual infrastructure elements, we focus on the city-scale land use and behavioral patterns, while addressing spatial heterogeneity and spatial autocorrelation. We implemented a spatially localized machine learning technique called Geographical Random Forest (GRF) on the California AV collision dataset. Analyzing multiple urban measures, including points of interest, building footprint, and land use, we built a GRF model and visualized it as a crash severity risk map of San Francisco. This paper presents three findings. First, spatially localized machine learning outperformed regular machine learning in predicting AV crash severity. The bias-variance tradeoff was evident as we adjusted the localization weight hyperparameter. Second, land use was the most important predictor, compared to intersections, building footprints, public transit stops, and Points Of Interest (POIs). Third, AV crashes were more likely to result in low-severity incidents in city center areas with greater diversity and commercial activities, than in residential neighborhoods. Residential land use is likely associated with higher severity due to human behavior and less restrictive environments. Counterintuitively, residential areas were associated with higher crash severity, compared to more complex areas such as commercial and mixed-use areas. When robotaxi operators train their AV systems, it is recommended to: (1) consider where their fleet operates and make localized algorithms for their perception system, and (2) design safety measures specific to residential neighborhoods, such as slower driving speeds and more alert sensors.

The demand for data privacy has led to the development of frameworks like Federated Graph Learning (FGL), which facilitate decentralized model training. However, a significant operational challenge in such systems is adhering to the right to be forgotten. This principle necessitates robust mechanisms for two distinct types of data removal: the selective erasure of specific entities and their associated knowledge from local subgraphs and the wholesale removal of a user's entire dataset and influence. Existing methods often struggle to fully address both unlearning requirements, frequently resulting in incomplete data removal or the persistence of residual knowledge within the system. This work introduces a unified framework, conceived to provide a comprehensive solution to these challenges. The proposed framework employs a bifurcated strategy tailored to the specific unlearning request. For fine-grained Meta Unlearning, it uses prototype gradients to direct the initial local forgetting process, which is then refined by generating adversarial graphs to eliminate any remaining data traces among affected clients. In the case of complete client unlearning, the framework utilizes adversarial graph generation exclusively to purge the departed client's contributions from the remaining network. Extensive experiments on multiple benchmark datasets validate the proposed approach. The framework achieves substantial improvements in model prediction accuracy across both client and meta-unlearning scenarios when compared to existing methods. Furthermore, additional studies confirm its utility as a plug-in module, where it materially enhances the predictive capabilities and unlearning effectiveness of other established methods.

Measuring innovation often relies on context-specific proxies and on expert evaluation. Hence, empirical innovation research is often limited to settings where such data is available. We investigate how large language models (LLMs) can be leveraged to overcome the constraints of manual expert evaluations and assist researchers in measuring innovation. We design an LLM framework that reliably approximates domain experts' assessment of innovation from unstructured text data. We demonstrate the performance and broad applicability of this framework through two studies in different contexts: (1) the innovativeness of software application updates and (2) the originality of user-generated feedback and improvement ideas in product reviews. We compared the performance (F1-score) and reliability (consistency rate) of our LLM framework against alternative measures used in prior innovation studies, and to state-of-the-art machine learning- and deep learning-based models. The LLM framework achieved higher F1-scores than the other approaches, and its results are highly consistent (i.e., results do not change across runs). This article equips R&D personnel in firms, as well as researchers, reviewers, and editors, with the knowledge and tools to effectively use LLMs for measuring innovation and evaluating the performance of LLM-based innovation measures. In doing so, we discuss, the impact of important design decisions-including model selection, prompt engineering, training data size, training data distribution, and parameter settings-on performance and reliability. Given the challenges inherent in using human expert evaluation and existing text-based measures, our framework has important implications for harnessing LLMs as reliable, increasingly accessible, and broadly applicable research tools for measuring innovation.

Low-rank adaptation (LoRA) has become a standard tool for efficiently fine-tuning large language models (LLMs). Yet, even minor LoRA updates can induce alignment drift, weakening safety and behavioral constraints through entangled parameter changes. To address this, we propose AlignGuard-LoRA (AGL), a principled framework for preserving alignment during finetuning. AGL introduces several key components: a primary task loss for supervision, Fisher Information Matrix-based regularization to restrict updates in alignment-sensitive subspaces, and task-specific regularization to stabilize the integration of new knowledge. We further introduce collision-aware regularization, blending Riemannian overlap -- which penalizes coordinate-wise interference -- and geodesic separation -- which encourages disjoint update geometry. We curate DriftCaps, a targeted diagnostic benchmark of safe and unsafe prompts designed to quantify alignment drift and safety degradation. Empirical evaluations show that AGL mitigates alignment drift by up to 50% on safety-critical benchmarks without degrading downstream task performance. Comprehensive ablation confirms that each component contributes distinctly to preserving latent safety behaviors. Finally, we derive and validate a scaling law for catastrophic forgetting, revealing that AGL flattens post-finetuning loss escalation while preserving adaptation dynamics. AGL is a structurally grounded refinement of LoRA, ensuring alignment preservation with minimal trade-offs. To encourage further exploration and development, we open-source our implementation.

Large Language Models (LLMs) fine-tuned to align with human values often exhibit alignment drift, producing unsafe or policy-violating completions when exposed to adversarial prompts, decoding perturbations, or paraphrased jailbreaks. While prior work has behaviorally characterized alignment failure, little is known about the training-time belief sources underlying these failures. We introduce TraceAlign, a unified framework for tracing unsafe completions back to their root causes in the model's training corpus. Central to our approach is the Belief Conflict Index (BCI), which quantifies semantic inconsistency between generated spans and aligned policies, based on retrieved training documents using suffix-array matching. We propose three complementary interventions: (i) TraceShield, an inference-time safety filter that refuses completions with high-BCI spans, (ii) Contrastive Belief Deconfliction Loss, a contrastive fine-tuning objective penalizing high-BCI continuations during DPO, and (iii) Prov-Decode, a provenance-aware decoding strategy that vetoes beam expansions predicted to yield high-BCI spans. Together, these defenses reduce alignment drift by up to 85% on our curated Alignment Drift Benchmark (ADB) while preserving utility on standard tasks, with delta less than 0.2 and improved refusal quality. We further derive a theoretical upper bound on drift likelihood via suffix-array span statistics, linking memorization frequency and length to adversarial reactivation risk. TraceAlign thus provides the first scalable, traceable, and grounded toolkit for understanding and mitigating alignment failures at source. To encourage further exploration and development, we open-source our implementation at: https://anonymous.4open.science/r/tracealign-2DA7

Graph unlearning is tailored for GNNs to handle widespread and various graph structure unlearning requests, which remain largely unexplored. The GIF (graph influence function) achieves validity under partial edge unlearning, but faces challenges in dealing with more disturbing node unlearning. To avoid the overhead of retraining and realize the model utility of unlearning, we proposed a novel node unlearning method to reverse the process of aggregation in GNN by embedding reconstruction and to adopt Range-Null Space Decomposition for the nodes' interaction learning. Experimental results on multiple representative datasets demonstrate the SOTA performance of our proposed approach.

Model merging via Mixture-of-Experts (MoE) has emerged as a scalable solution for consolidating multiple task-specific models into a unified sparse architecture, where each expert is derived from a model fine-tuned on a distinct task. While effective for multi-task integration, this paradigm introduces a critical yet underexplored challenge: how to attribute and protect the intellectual property (IP) of individual experts after merging. We propose RouteMark, a framework for IP protection in merged MoE models through the design of expert routing fingerprints. Our key insight is that task-specific experts exhibit stable and distinctive routing behaviors under probing inputs. To capture these patterns, we construct expert-level fingerprints using two complementary statistics: the Routing Score Fingerprint (RSF), quantifying the intensity of expert activation, and the Routing Preference Fingerprint (RPF), characterizing the input distribution that preferentially activates each expert. These fingerprints are reproducible, task-discriminative, and lightweight to construct. For attribution and tampering detection, we introduce a similarity-based matching algorithm that compares expert fingerprints between a suspect and a reference (victim) model. Extensive experiments across diverse tasks and CLIP-based MoE architectures show that RouteMark consistently yields high similarity for reused experts and clear separation from unrelated ones. Moreover, it remains robust against both structural tampering (expert replacement, addition, deletion) and parametric tampering (fine-tuning, pruning, permutation), outperforming weight- and activation-based baseliness. Our work lays the foundation for RouteMark as a practical and broadly applicable framework for IP verification in MoE-based model merging.