搜索结果：behaviors

Supervised fine-tuning (SFT) induces new behaviors in large language models, yet imposes no structural constraint on how these behaviors are distributed within the model. Existing behavior interpretation methods, such as circuit attribution approaches, identify sparse subnetworks correlated with SFT-induced behaviors post-hoc. However, such correlations do not imply *causal necessity*, limiting the ability to selectively control SFT-induced behaviors at inference time. We pursue an alternative by asking: can an SFT-induced behavior be deliberately compressed into a sparse, mechanistically necessary subnetwork, termed a *carrier*, while remaining controllable at inference time without weight modification? We propose (a) **Loss-Constrained Dual Descent (LCDD)**, which constructs such carriers by jointly optimizing routing masks and model weights under an explicit utility budget, and (b) **SFT-Eraser**, a soft prompt optimized via activation matching on extracted carrier channels, to reverse the SFT-induced behavior. Across safety, fixed-response, and style behaviors on multiple model families, LCDD yields sparse carriers that preserve target behaviors while enabling strong reversion

Timestamps in digital traces include significant detailed information on when human behaviors occur, which is universally available and standardized in all types of digital traces. Nevertheless, the concept of time is under-explicated in empirical studies of online behaviors. This paper discusses the (un)desirable properties of timestamps in digital traces and summarizes how timestamps in digital traces have been utilized in existing studies of online behaviors. The paper argues that time-in-behaviors perspective can provide a microscope with a renovated temporal lens to observe and understand online behaviors. Going beyond the traditional behaviors-in-time perspective, time-in-behaviors perspective enables empirical examination of online behaviors from multiple units of analysis (e.g., discrete behaviors, behavioral sessions, and behavioral trajectories) and from multiple dimensions (e.g., duration, order, transition, rhythm). The paper shows the potentials of the time-in-behaviors perspective with several empirical cases and proposes future directions in explicating the concept of time in computational social science.

Imitation Learning (IL) techniques aim to replicate human behaviors in specific tasks. While IL has gained prominence due to its effectiveness and efficiency, traditional methods often focus on datasets collected from experts to produce a single efficient policy. Recently, extensions have been proposed to handle datasets of diverse behaviors by mainly focusing on learning transition-level diverse policies or on performing entropy maximization at the trajectory level. While these methods may lead to diverse behaviors, they may not be sufficient to reproduce the actual diversity of demonstrations or to allow controlled trajectory generation. To overcome these drawbacks, we propose a different method based on two key features: a) Temporal Consistency that ensures consistent behaviors across entire episodes and not just at the transition level as well as b) Controllability obtained by constructing a latent space of behaviors that allows users to selectively activate specific behaviors based on their requirements. We compare our approach to state-of-the-art methods over a diverse set of tasks and environments. Project page: https://mathieu-petitbois.github.io/projects/swr/

Predicting user behavior is essential for intelligent assistant services, yet deep learning models often struggle to capture long-tailed behaviors. Large language models (LLMs), with their pretraining on vast corpora containing rich behavioral knowledge, offer promise. However, existing fine-tuning approaches tend to overfit to frequent ``anchor'' behaviors, reducing their ability to predict less common ``tail'' behaviors. In this paper, we introduce BehaviorLM, a progressive fine-tuning approach that addresses this issue. In the first stage, LLMs are fine-tuned on anchor behaviors while preserving general behavioral knowledge. In the second stage, fine-tuning uses a balanced subset of all behaviors based on sample difficulty to improve tail behavior predictions without sacrificing anchor performance. Experimental results on two real-world datasets demonstrate that BehaviorLM robustly predicts both anchor and tail behaviors and effectively leverages LLM behavioral knowledge to master tail behavior prediction with few-shot examples.

Learning highly dynamic behaviors for robots has been a longstanding challenge. Traditional approaches have demonstrated robust locomotion, but the exhibited behaviors lack diversity and agility. They employ approximate models, which lead to compromises in performance. Data-driven approaches have been shown to reproduce agile behaviors of animals, but typically have not been able to learn highly dynamic behaviors. In this paper, we propose a learning-based approach to enable robots to learn highly dynamic behaviors from animal motion data. The learned controller is deployed on a quadrupedal robot and the results show that the controller is able to reproduce highly dynamic behaviors including sprinting, jumping and sharp turning. Various behaviors can be activated through human interaction using a stick with markers attached to it. Based on the motion pattern of the stick, the robot exhibits walking, running, sitting and jumping, much like the way humans interact with a pet.

Anthropomorphism in artificial intelligence (AI) is a growing area of interest, as AI systems increasingly exhibit human-like expressions, behaviors and interaction styles. This research serves as a systematic observation and categorization of anthropomorphic behaviors in AI outputs. Anthropomorphic behavior refers to the deliberate or emergent manifestation of human-like expressions or linguistic cues in system outputs, such as demonstrating empathy. Using a behaviorally driven taxonomy, our study identifies key forms of anthropomorphic behaviors in the responses of ChatGPT and examines their implications for the theory, practice and ethics of AI systems. The taxonomy enables more nuanced detection of anthropomorphism, offering value to developers and policymakers in balancing the benefits with the potential risks. This work contributes to the academic discourse by providing a foundation for future efforts to refine, expand, and automate the detection of anthropomorphic behavior across diverse AI applications.

Multi-behavior recommendation faces a critical challenge in practice: auxiliary behaviors (e.g., clicks, carts) are often noisy, weakly correlated, or semantically misaligned with the target behavior (e.g., purchase), which leads to biased preference learning and suboptimal performance. While existing methods attempt to fuse these heterogeneous signals, they inherently lack a principled mechanism to ensure robustness against such behavioral inconsistency. In this work, we propose Robust Multi-Behavior Recommendation towards Target Behaviors (RMBRec), a robust multi-behavior recommendation framework grounded in an information-theoretic robustness principle. We interpret robustness as a joint process of maximizing predictive information while minimizing its variance across heterogeneous behavioral environments. Under this perspective, the Representation Robustness Module (RRM) enhances local semantic consistency by maximizing the mutual information between users' auxiliary and target representations, whereas the Optimization Robustness Module (ORM) enforces global stability by minimizing the variance of predictive risks across behaviors, which is an efficient approximation to invaria

The Android ecosystem faces a notable challenge known as fragmentation, which denotes the extensive diversity within the system. This issue is mainly related to differences in system versions, device hardware specifications, and customizations introduced by manufacturers. The growing divergence among devices leads to marked variations in how a given app behaves across diverse devices. This is referred to as device-specific behaviors. In this work, we present the first large-scale empirical study of device-specific behaviors in real-world Android apps. We have designed a three-phase static analysis framework to accurately detect and understand the device-specific behaviors. Upon employing our tool on a dataset comprising more than 20,000 apps, we detected device-specific behaviors in 2,357 of them. By examining the distribution of device-specific behaviors, our analysis revealed that apps within the Chinese third-party app market exhibit more relevant behaviors compared to their counterparts in Google Play. Additionally, these behaviors are more likely to feature dominant brands that hold larger market shares. Reflecting this, we have classified these device-specific behaviors into

Feint behaviors refer to a set of deceptive behaviors in a nuanced manner, which enable players to obtain temporal and spatial advantages over opponents in competitive games. Such behaviors are crucial tactics in most competitive multi-player games (e.g., boxing, fencing, basketball, motor racing, etc.). However, existing literature does not provide a comprehensive (and/or concrete) formalization for Feint behaviors, and their implications on game strategies. In this work, we introduce the first comprehensive formalization of Feint behaviors at both action-level and strategy-level, and provide concrete implementation and quantitative evaluation of them in multi-player games. The key idea of our work is to (1) allow automatic generation of Feint behaviors via Palindrome-directed templates, combine them into meaningful behavior sequences via a Dual-Behavior Model; (2) concertize the implications from our formalization of Feint on game strategies, in terms of temporal, spatial and their collective impacts respectively; and (3) provide a unified implementation scheme of Feint behaviors in existing MARL frameworks. The experimental results show that our design of Feint behaviors can (1)

Effective human-AI coordination requires artificial agents capable of exhibiting and responding to human-like behaviors while adapting to changing contexts. Imitation learning has emerged as one of the prominent approaches to build such agents by training them to mimic human-demonstrated behaviors. However, current methods struggle to capture the inherent diversity and non-Markovian nature of human behavior and lack the ability to steer behavior at inference time. Drawing inspiration from the theory of human cognitive processes, where inner speech guides action selection before execution, we propose MIMIC (Modeling Inner Motivations for Imitation and Control), a framework that uses language as an internal representation of behavioral intent. MIMIC employs the novel use of vision-language models as linguistic scaffolding to train a conditional variational autoencoder capable of generating inner speech from observations. A diffusion-based behavior cloning policy then selects actions conditioned on current observations and the generated inner speech. MIMIC enables fine-grained steering of behavior at inference time by conditioning the agent on behavior-specific speech. Experiments acr

Imitation learning (IL) has shown promise in various applications (e.g. robot locomotion) but is often limited to learning a single expert policy, constraining behavior diversity and robustness in unpredictable real-world scenarios. To address this, we introduce Quality Diversity Inverse Reinforcement Learning (QD-IRL), a novel framework that integrates quality-diversity optimization with IRL methods, enabling agents to learn diverse behaviors from limited demonstrations. This work introduces Extrinsic Behavioral Curiosity (EBC), which allows agents to receive additional curiosity rewards from an external critic based on how novel the behaviors are with respect to a large behavioral archive. To validate the effectiveness of EBC in exploring diverse locomotion behaviors, we evaluate our method on multiple robot locomotion tasks. EBC improves the performance of QD-IRL instances with GAIL, VAIL, and DiffAIL across all included environments by up to 185%, 42%, and 150%, even surpassing expert performance by 20% in Humanoid. Furthermore, we demonstrate that EBC is applicable to Gradient-Arborescence-based Quality Diversity Reinforcement Learning (QD-RL) algorithms, where it substantiall

We study the universal dynamical relaxation behaviors of a quantum XY chain following a quench, paying special attention to the case that the prequenched Hamiltonian, or the postquenched Hamiltonian, or both of them are at critical points of equilibrium quantum phase transitions. In such ``critical quench", we find very interesting real-time dynamical scaling behaviors and the crossover phenomena between them. For a quench from a noncritical point to a critical point, we find that, compared to the noncritical quench, the universal power-law scaling behavior does not change; however, there may be a crossover between the exponential decaying behavior and the power-law scaling. For a quench from a critical point to a noncritical point, the power-law scaling behaviors $t^{-3/2}$ and $t^{-3/4}$ in the noncritical quenches may be changed to $t^{-1}$ and $t^{-1/2}$, respectively. If the prequenched Hamiltonian is set to be a point that is close to but not exactly at a critical point, we find interesting crossover phenomena between different power-law scaling behaviors. We also study the quench from the vicinity of a multicritical point, we find crossover behaviors that are induced by a di

Modelling other agents' behaviors plays an important role in decision models for interactions among multiple agents. To optimise its own decisions, a subject agent needs to model what other agents act simultaneously in an uncertain environment. However, modelling insufficiency occurs when the agents are competitive and the subject agent can not get full knowledge about other agents. Even when the agents are collaborative, they may not share their true behaviors due to their privacy concerns. In this article, we investigate into diversifying behaviors of other agents in the subject agent's decision model prior to their interactions. Starting with prior knowledge about other agents' behaviors, we use a linear reduction technique to extract representative behavioral features from the known behaviors. We subsequently generate their new behaviors by expanding the features and propose two diversity measurements to select top-K behaviors. We demonstrate the performance of the new techniques in two well-studied problem domains. This research will contribute to intelligent systems dealing with unknown unknowns in an open artificial intelligence world.

Understanding the principles of real-world biological multi-agent behaviors is a current challenge in various scientific and engineering fields. The rules regarding the real-world biological multi-agent behaviors such as team sports are often largely unknown due to their inherently higher-order interactions, cognition, and body dynamics. Estimation of the rules from data, i.e., data-driven approaches such as machine learning, provides an effective way for the analysis of such behaviors. Although most data-driven models have non-linear structures and high prediction performances, it is sometimes hard to interpret them. This survey focuses on data-driven analysis for quantitative understanding of invasion team sports behaviors such as basketball and football, and introduces two main approaches for understanding such multi-agent behaviors: (1) extracting easily interpretable features or rules from data and (2) generating and controlling behaviors in visually-understandable ways. The first approach involves the visualization of learned representations and the extraction of mathematical structures behind the behaviors. The second approach can be used to test hypotheses by simulating and

Evolving swarm behaviors with artificial agents is computationally expensive and challenging. Because reward structures are often sparse in swarm problems, only a few simulations among hundreds evolve successful swarm behaviors. Additionally, swarm evolutionary algorithms typically rely on ad hoc fitness structures, and novel fitness functions need to be designed for each swarm task. This paper evolves swarm behaviors by systematically combining Postcondition-Precondition-Action (PPA) canonical Behavior Trees (BT) with a Grammatical Evolution. The PPA structure replaces ad hoc reward structures with systematic postcondition checks, which allows a common grammar to learn solutions to different tasks using only environmental cues and BT feedback. The static performance of learned behaviors is poor because no agent learns all necessary subtasks, but performance while evolving is excellent because agents can quickly change behaviors in new contexts. The evolving algorithm succeeded in 75\% of learning trials for both foraging and nest maintenance tasks, an eight-fold improvement over prior work.