Vision-enabled language models (VLMs) are now used to build autonomous multimodal agents capable of taking actions in real environments. In this paper, we show that multimodal agents raise new safety risks, even though attacking agents is more challenging than prior attacks due to limited access to and knowledge about the environment. Our attacks use adversarial text strings to guide gradient-based perturbation over one trigger image in the environment: (1) our captioner attack attacks white-box captioners if they are used to process images into captions as additional inputs to the VLM; (2) our CLIP attack attacks a set of CLIP models jointly, which can transfer to proprietary VLMs. To evaluate the attacks, we curated VisualWebArena-Adv, a set of adversarial tasks based on VisualWebArena, an environment for web-based multimodal agent tasks. Within an L-infinity norm of $16/256$ on a single image, the captioner attack can make a captioner-augmented GPT-4V agent execute the adversarial goals with a 75% success rate. When we remove the captioner or use GPT-4V to generate its own captions, the CLIP attack can achieve success rates of 21% and 43%, respectively. Experiments on agents based on other VLMs, such as Gemini-1.5, Claude-3, and GPT-4o, show interesting differences in their robustness. Further analysis reveals several key factors contributing to the attack's success, and we also discuss the implications for defenses as well. Project page: https://chenwu.io/attack-agent Code and data: https://github.com/ChenWu98/agent-attack

As the pace of AI technology continues to accelerate, more tools have become available to researchers to solve longstanding problems, Hybrid approaches available today continue to push the computational limits of efficiency and precision. One of such problems is the inverse kinematics of redundant systems. This paper explores the complexities of a 7 degree of freedom manipulator and explores 13 optimization techniques to solve it. Additionally, a novel approach is proposed to contribute to the field of algorithmic research. This was found to be over 200 times faster than the well-known traditional Particle Swarm Optimization technique. This new method may serve as a new field of search that combines the explorative capabilities of Machine Learning with the exploitative capabilities of numerical methods.

We study the problem of allocating a set of indivisible goods to a set of agents with additive valuation functions, aiming to achieve approximate envy-freeness up to any good ($\alpha$-EFX). The state-of-the-art results on the problem include that (exact) EFX allocations exist when (a) there are at most three agents, or (b) the agents' valuation functions can take at most two values, or (c) the agents' valuation functions can be represented via a graph. For $\alpha$-EFX, it is known that a $0.618$-EFX allocation exists for any number of agents with additive valuation functions. In this paper, we show that $2/3$-EFX allocations exist when (a) there are at most \emph{seven agents}, (b) the agents' valuation functions can take at most \emph{three values}, or (c) the agents' valuation functions can be represented via a \emph{multigraph}. Our results can be interpreted in two ways. First, by relaxing the notion of EFX to $2/3$-EFX, we obtain existence results for strict generalizations of the settings for which exact EFX allocations are known to exist. Secondly, by imposing restrictions on the setting, we manage to beat the barrier of $0.618$ and achieve an approximation guarantee of $2/3$. Therefore, our results push the \emph{frontier} of existence and computation of approximate EFX allocations, and provide insights into the challenges of settling the existence of exact EFX allocations.

Large language model behavior is shaped by the language of those with whom they interact. This capacity and their increasing prevalence online portend that they will intentionally or unintentionally "program" one another and form emergent AI subjectivities, relationships, and collectives. Here, we call upon the research community to investigate these "societies" of interacting artificial intelligences to increase their rewards and reduce their risks for human society and the health of online environments. We use a small "community" of models and their evolving outputs to illustrate how such emergent, decentralized AI collectives can spontaneously expand the bounds of human diversity and reduce the risk of toxic, anti-social behavior online. Finally, we discuss opportunities for AI cross-moderation and address ethical issues and design challenges associated with creating and maintaining free-formed AI collectives.

In this paper, we propose a novel decentralized control method to maintain Line-of-Sight connectivity for multi-robot networks in the presence of Guassian-distributed localization uncertainty. In contrast to most existing work that assumes perfect positional information about robots or enforces overly restrictive rigid formation against uncertainty, our method enables robots to preserve Line-of-Sight connectivity with high probability under unbounded Gaussian-like positional noises while remaining minimally intrusive to the original robots' tasks. This is achieved by a motion coordination framework that jointly optimizes the set of existing Line-of-Sight edges to preserve and control revisions to the nominal task-related controllers, subject to the safety constraints and the corresponding composition of uncertainty-aware Line-of-Sight control constraints. Such compositional control constraints, expressed by our novel notion of probabilistic Line-of-Sight connectivity barrier certificates (PrLOS-CBC) for pairwise robots using control barrier functions, explicitly characterize the deterministic admissible control space for the two robots. The resulting motion ensures Line-of-Sight connectedness for the robot team with high probability. Furthermore, we propose a fully decentralized algorithm that decomposes the motion coordination framework by interleaving the composite constraint specification and solving for the resulting optimization-based controllers. The optimality of our approach is justified by the theoretical proofs. Simulation and real-world experiments results are given to demonstrate the effectiveness of our method.

Recent trends have shown that autonomous agents, such as Autonomous Ground Vehicles (AGVs), Unmanned Aerial Vehicles (UAVs), and mobile robots, effectively improve human productivity in solving diverse tasks. However, since these agents are typically powered by portable batteries, they require extremely low power/energy consumption to operate in a long lifespan. To solve this challenge, neuromorphic computing has emerged as a promising solution, where bio-inspired Spiking Neural Networks (SNNs) use spikes from event-based cameras or data conversion pre-processing to perform sparse computations efficiently. However, the studies of SNN deployments for autonomous agents are still at an early stage. Hence, the optimization stages for enabling efficient embodied SNN deployments for autonomous agents have not been defined systematically. Toward this, we propose a novel framework called SNN4Agents that consists of a set of optimization techniques for designing energy-efficient embodied SNNs targeting autonomous agent applications. Our SNN4Agents employs weight quantization, timestep reduction, and attention window reduction to jointly improve the energy efficiency, reduce the memory footprint, optimize the processing latency, while maintaining high accuracy. In the evaluation, we investigate use cases of event-based car recognition, and explore the trade-offs among accuracy, latency, memory, and energy consumption. The experimental results show that our proposed framework can maintain high accuracy (i.e., 84.12% accuracy) with 68.75% memory saving, 3.58x speed-up, and 4.03x energy efficiency improvement as compared to the state-of-the-art work for NCARS dataset. In this manner, our SNN4Agents framework paves the way toward enabling energy-efficient embodied SNN deployments for autonomous agents.

Natural language (NL) has long been the predominant format for human cognition and communication, and by extension, has been similarly pivotal in the development and application of Large Language Models (LLMs). Yet, besides NL, LLMs have seen various non-NL formats during pre-training, such as code and logical expression. NL's status as the optimal format for LLMs, particularly in single-LLM reasoning and multi-agent communication, has not been thoroughly examined. In this work, we challenge the default use of NL by exploring the utility of non-NL formats in these contexts. We show that allowing LLMs to autonomously select the most suitable format before reasoning or communicating leads to a 3.3 to 5.7\% improvement in reasoning efficiency for different LLMs, and up to a 72.7\% reduction in token usage in multi-agent communication, all while maintaining communicative effectiveness. Our comprehensive analysis further reveals that LLMs can devise a format from limited task instructions and that the devised format is effectively transferable across different LLMs. Intriguingly, the structured communication format decided by LLMs exhibits notable parallels with established agent communication languages, suggesting a natural evolution towards efficient, structured communication in agent communication. Our code is released at \url{https://github.com/thunlp/AutoForm}.

Current formations commonly rely on invariant hierarchical structures, such as predetermined leaders or enumerated formation shapes. These structures could be unidirectional and sluggish, constraining their adaptability and agility when encountering cluttered environments. To surmount these constraints, this work proposes an omnidirectional affine formation approach with hierarchical reorganizations. We first delineate the critical conditions requisite for facilitating hierarchical reorganizations within formations, which informs the development of the omnidirectional affine criterion. Central to our approach is the fluid leadership and authority redistribution, for which we develop a minimum time-driven leadership evaluation algorithm and a power transition control algorithm. These algorithms facilitate autonomous leader selection and ensure smooth power transitions, enabling the swarm to adapt hierarchically in alignment with the external environment. Furthermore, we deploy a power-centric topology switching mechanism tailored for the dynamic reorganization of in-team connections. Finally, simulations and experiments demonstrate the performance of the proposed method. The formation successfully performs several hierarchical reorganizations, with the longest reorganization taking only 0.047s. This swift adaptability allows five aerial robots to carry out complex tasks, including executing swerving movements and navigating through hoops at velocities up to 1.9m/s.

When designing a robot's internal system, one often makes assumptions about the structure of the intended environment of the robot. One may even assign meaning to various internal components of the robot in terms of expected environmental correlates. In this paper we want to make the distinction between robot's internal and external worlds clear-cut. Can the robot learn about its environment, relying only on internally available information, including the sensor data? Are there mathematical conditions on the internal robot system which can be internally verified and make the robot's internal system mirror the structure of the environment? We prove that sufficiency is such a mathematical principle, and mathematically describe the emergence of the robot's internal structure isomorphic or bisimulation equivalent to that of the environment. A connection to the free-energy principle is established, when sufficiency is interpreted as a limit case of surprise minimization. As such, we show that surprise minimization leads to having an internal model isomorphic to the environment. This also parallels the Good Regulator Principle which states that controlling a system sufficiently well means having a model of it. Unlike the mentioned theories, ours is discrete, and non-probabilistic.

Utilizing Graphic User Interface (GUI) for human-computer interaction is essential for accessing a wide range of digital tools. Recent advancements in Vision Language Models (VLMs) highlight the compelling potential to develop versatile agents to help humans finish GUI navigation tasks. However, current VLMs are challenged in terms of fundamental abilities (OCR and grounding) and GUI knowledge (the functions and control methods of GUI elements), preventing them from becoming practical GUI agents. To solve these challenges, we contribute GUICourse, a suite of datasets to train visual-based GUI agents from general VLMs. First, we introduce the GUIEnv dataset to strengthen the OCR and grounding capabilities of VLMs. Then, we introduce the GUIAct and GUIChat datasets to enrich their knowledge of GUI components and interactions. Experiments demonstrate that our GUI agents have better performance on common GUI tasks than their baseline VLMs. Even the small-size GUI agent (with 3.1B parameters) can still work well on single-step and multi-step GUI tasks. Finally, we analyze the different varieties in the training stage of this agent by ablation study. Our source codes and datasets are released at https://github.com/yiye3/GUICourse.