Recent advances in large language models (LLMs) have enabled a new class of AI agents that automate multiple stages of the data science workflow by integrating planning, tool use, and multimodal reasoning across text, code, tables, and visuals. This survey presents the first comprehensive, lifecycle-aligned taxonomy of data science agents, systematically analyzing and mapping forty-five systems onto the six stages of the end-to-end data science process: business understanding and data acquisition, exploratory analysis and visualization, feature engineering, model building and selection, interpretation and explanation, and deployment and monitoring. In addition to lifecycle coverage, we annotate each agent along five cross-cutting design dimensions: reasoning and planning style, modality integration, tool orchestration depth, learning and alignment methods, and trust, safety, and governance mechanisms. Beyond classification, we provide a critical synthesis of agent capabilities, highlight strengths and limitations at each stage, and review emerging benchmarks and evaluation practices. Our analysis identifies three key trends: most systems emphasize exploratory analysis, visualization, and modeling while neglecting business understanding, deployment, and monitoring; multimodal reasoning and tool orchestration remain unresolved challenges; and over 90% lack explicit trust and safety mechanisms. We conclude by outlining open challenges in alignment stability, explainability, governance, and robust evaluation frameworks, and propose future research directions to guide the development of robust, trustworthy, low-latency, transparent, and broadly accessible data science agents.

Large language models (LLMs) are increasingly deployed in agentic systems where they map user intents to relevant external tools to fulfill a task. A critical step in this process is tool selection, where a retriever first surfaces candidate tools from a larger pool, after which the LLM selects the most appropriate one. This pipeline presents an underexplored attack surface where errors in selection can lead to severe outcomes like unauthorized data access or denial of service, all without modifying the agent's model or code. While existing evaluations measure task performance in benign settings, they overlook the specific vulnerabilities of the tool selection mechanism under adversarial conditions. To address this gap, we introduce ToolCert, the first statistical framework that formally certifies tool selection robustness. ToolCert models tool selection as a Bernoulli success process and evaluates it against a strong, adaptive attacker who introduces adversarial tools with misleading metadata, and are iteratively refined based on the agent's previous choices. By sampling these adversarial interactions, ToolCert produces a high-confidence lower bound on accuracy, formally quantifying the agent's worst-case performance. Our evaluation with ToolCert uncovers the severe fragility: under attacks injecting deceptive tools or saturating retrieval, the certified accuracy bound drops near zero, an average performance drop of over 60% compared to non-adversarial settings. For attacks targeting the retrieval and selection stages, the certified accuracy bound plummets to less than 20% after just a single round of adversarial adaptation. ToolCert thus reveals previously unexamined security threats inherent to tool selection and provides a principled method to quantify an agent's robustness to such threats, a necessary step for the safe deployment of agentic systems.

Personalized AI agents are becoming central to modern information retrieval, yet most evaluation methodologies remain static, relying on fixed benchmarks and one-off metrics that fail to reflect how users' needs evolve over time. These limitations hinder our ability to assess whether agents can meaningfully adapt to individuals across dynamic, longitudinal interactions. In this perspective paper, we propose a conceptual lens for rethinking evaluation in adaptive personalization, shifting the focus from static performance snapshots to interaction-aware, evolving assessments. We organize this lens around three core components: (1) persona-based user simulation with temporally evolving preference models; (2) structured elicitation protocols inspired by reference interviews to extract preferences in context; and (3) adaptation-aware evaluation mechanisms that measure how agent behavior improves across sessions and tasks. While recent works have embraced LLM-driven user simulation, we situate this practice within a broader paradigm for evaluating agents over time. To illustrate our ideas, we conduct a case study in e-commerce search using the PersonalWAB dataset. Beyond presenting a framework, our work lays a conceptual foundation for understanding and evaluating personalization as a continuous, user-centric endeavor.

Strategy logic (SL) is a powerful temporal logic that enables first-class reasoning over strategic behavior in multi-agent systems (MAS). In many MASs, the agents (and their strategies) cannot observe the global state of the system, leading to many extensions of SL centered around imperfect information, such as strategy logic with imperfect information (SL$_\mathit{ii}$). Along orthogonal lines, researchers have studied the combination of strategic behavior and hyperproperties. Hyperproperties are system properties that relate multiple executions in a system and commonly arise when specifying security policies. Hyper Strategy Logic (HyperSL) is a temporal logic that combines quantification over strategies with the ability to express hyperproperties on the executions of different strategy profiles. In this paper, we study the relation between SL$_\mathit{ii}$ and HyperSL. Our main result is that both logics (restricted to formulas where no state formulas are nested within path formulas) are equivalent in the sense that we can encode SL$_\mathit{ii}$ instances into HyperSL instances and vice versa. For the former direction, we build on the well-known observation that imperfect information is a hyperproperty. For the latter direction, we construct a self-composition of MASs and show how we can simulate hyperproperties using imperfect information.

This position paper presents A4FN, an Agentic Artificial Intelligence (AI) architecture for intent-driven automation in Flying Networks (FNs) using Unmanned Aerial Vehicles (UAVs) as access nodes. A4FN leverages Generative AI and Large Language Models (LLMs) to enable real-time, context-aware network control via a distributed agentic system. It comprises two components: the Perception Agent (PA), which semantically interprets multimodal input -- including imagery, audio, and telemetry data -- from UAV-mounted sensors to derive Service Level Specifications (SLSs); and the Decision-and-Action Agent (DAA), which reconfigures the network based on inferred intents. A4FN embodies key properties of Agentic AI, including autonomy, goal-driven reasoning, and continuous perception-action cycles. Designed for mission-critical, infrastructure-limited scenarios such as disaster response, it supports adaptive reconfiguration, dynamic resource management, and interoperability with emerging wireless technologies. The paper details the A4FN architecture, its core innovations, and open research challenges in multi-agent coordination and Agentic AI integration in next-generation FNs.

High Altitude Balloons (HABs) can leverage stratospheric wind layers for limited horizontal control, enabling applications in reconnaissance, environmental monitoring, and communications networks. Existing multi-agent HAB coordination approaches use deterministic methods like Voronoi partitioning and extremum seeking control for large global constellations, which perform poorly for smaller teams and localized missions. While single-agent HAB control using reinforcement learning has been demonstrated on HABs, coordinated multi-agent reinforcement learning (MARL) has not yet been investigated. This work presents the first systematic application of multi-agent reinforcement learning (MARL) to HAB coordination for distributed area coverage. We extend our previously developed reinforcement learning simulation environment (RLHAB) to support cooperative multi-agent learning, enabling multiple agents to operate simultaneously in realistic atmospheric conditions. We adapt QMIX for HAB area coverage coordination, leveraging Centralized Training with Decentralized Execution to address atmospheric vehicle coordination challenges. Our approach employs specialized observation spaces providing individual state, environmental context, and teammate data, with hierarchical rewards prioritizing coverage while encouraging spatial distribution. We demonstrate that QMIX achieves similar performance to the theoretically optimal geometric deterministic method for distributed area coverage, validating the MARL approach and providing a foundation for more complex autonomous multi-HAB missions where deterministic methods become intractable.

Robots and other intelligent systems navigating in complex dynamic environments should predict future actions and intentions of surrounding agents to reach their goals efficiently and avoid collisions. The dynamics of those agents strongly depends on their tasks, roles, or observable labels. Class-conditioned motion prediction is thus an appealing way to reduce forecast uncertainty and get more accurate predictions for heterogeneous agents. However, this is hardly explored in the prior art, especially for mobile robots and in limited data applications. In this paper, we analyse different class-conditioned trajectory prediction methods on two datasets. We propose a set of conditional pattern-based and efficient deep learning-based baselines, and evaluate their performance on robotics and outdoors datasets (THÖR-MAGNI and Stanford Drone Dataset). Our experiments show that all methods improve accuracy in most of the settings when considering class labels. More importantly, we observe that there are significant differences when learning from imbalanced datasets, or in new environments where sufficient data is not available. In particular, we find that deep learning methods perform better on balanced datasets, but in applications with limited data, e.g., cold start of a robot in a new environment, or imbalanced classes, pattern-based methods may be preferable.

The emergence of cooperation in the groups of interacting agents is one of the most fascinating phenomena observed in many complex systems studied in social science and ecology, even in the situations where one would expect the agent to use a free-rider policy. This is especially surprising in the situation where no external mechanisms based on reputation or punishment are present. One of the possible explanations of this effect is the inhomogeneity of the various aspects of interactions, which can be used to clarify the seemingly paradoxical behavior. In this report we demonstrate that the diversity of interaction networks helps to some degree to explain the emergence of cooperation. We extend the model of spatial interaction diversity introduced in [L. Shang et al., Physica A, 593:126999 (2022)] by enabling the evaluation of the interaction groups. We show that the process of the reevaluation of the interaction group facilitates the emergence of cooperation. Furthermore, we also observe that a significant participation of agents switching their interaction neighborhoods has a negative impact on the formation of cooperation. The introduced scenario can help to understand the formation of cooperation in the systems where no additional mechanisms for controlling agents are included.

Deploying capable and user-aligned LLM-based systems necessitates reliable evaluation. While LLMs excel in verifiable tasks like coding and mathematics, where gold-standard solutions are available, adoption remains challenging for subjective tasks that lack a single correct answer. E-commerce Query Rewriting (QR) is one such problem where determining whether a rewritten query properly captures the user intent is extremely difficult to figure out algorithmically. In this work, we introduce OptAgent, a novel framework that combines multi-agent simulations with genetic algorithms to verify and optimize queries for QR. Instead of relying on a static reward model or a single LLM judge, our approach uses multiple LLM-based agents, each acting as a simulated shopping customer, as a dynamic reward signal. The average of these agent-derived scores serves as an effective fitness function for an evolutionary algorithm that iteratively refines the user's initial query. We evaluate OptAgent on a dataset of 1000 real-world e-commerce queries in five different categories, and we observe an average improvement of 21.98% over the original user query and 3.36% over a Best-of-N LLM rewriting baseline.

Evaluating the quality of multi-turn chatbot interactions remains challenging, as most existing methods assess interactions at the turn level without addressing whether a user's overarching goal was fulfilled. A ``goal'' here refers to an information need or task, such as asking for policy information or applying for leave. We propose a comprehensive framework for goal-oriented evaluation of multi-agent systems (MAS), introducing the \textbf{Goal Success Rate (GSR)} to measure the percentage of fulfilled goals, and a \textbf{Root Cause of Failure (RCOF)} taxonomy to identify reasons for failure in multi-agent chatbots. Our method segments conversations by user goals and evaluates success using all relevant turns. We present a model-based evaluation system combining teacher LLMs, where domain experts define goals, set quality standards serving as a guidance for the LLMs. The LLMs use ``thinking tokens'' to produce interpretable rationales, enabling \textit{explainable}, \textit{data-efficient} evaluations. In an enterprise setting, we apply our framework to evaluate AIDA, a zero-to-one employee conversational agent system built as a ground-up multi-agent conversational agent, and observe GSR improvement from 63\% to 79\% over six months since its inception. Our framework is generic and offers actionable insights through a detailed defect taxonomy based on analysis of failure points in multi-agent chatbots, diagnosing overall success, identifying key failure modes, and informing system improvements.