Intent Inference Research Articles

Long-Term and Short-Term Opponent Intention Inference for Football Multiplayer Policy Learning

Long-Term and Short-Term Opponent Intention Inference for Football Multiplayer Policy Learning

Frontal Cortex Acts as Causality Transition Hub from Mirror Network to Mentalizing Network During Action Intention Understanding.

Introduction: While understanding other's action intention, mirror and mentalizing systems of human brain are successively activated in action perception and intention inference processes. Methods: To reveal the relationship between mirror and mentalizing systems during the two stages, this electroencephalogram study adopted the method of time-varying orthogonalized partial directed coherence (OPDC) to assess causal interaction between mirror and mentalizing networks during a "hand-cup interaction" action intention understanding task. Results: Task-related causal connectivity was found in gamma frequency band (30-45 Hz), primarily manifested as directed edges from sensorimotor to frontal areas in poststimulus 400-600 ms interval and directed links from frontal to parietal and temporal regions in 600-800 ms period. The analysis of event-related potential and source currents suggests that the change of inter-regional causality is related with functional transition of the brain from mirror matching to intention inference. The OPDC network modeling further finds that frontal area contains more inflow nodes in mirror network, whereas more outflow nodes in mentalizing network, with high betweenness centrality in temporally changing functional communities. Compared with intention-oriented actions, identification of unintelligible action intention particularly induces stronger OPDC from right superior frontal to inferior frontal gyrus and from sensorimotor to right frontotemporal regions during mentalizing inference process. Conclusion: These findings collectively suggest that, in the time ordering of information transfer within the directed networks, frontal area plays an important role of bridging hub between mirror and mentalizing systems, from maintaining and supervising perceptual information for mirror matching to controlling the mentalizing process for decoding other's action intention.

SoftBoMI: a non-invasive wearable body-machine interface for mapping movement of shoulder to commands

Objective.Customized human-machine interfaces for controlling assistive devices are vital in improving the self-help ability of upper limb amputees and tetraplegic patients. Given that most of them possess residual shoulder mobility, using it to generate commands to operate assistive devices can serve as a complementary approach to brain-computer interfaces.Approach.We propose a hybrid body-machine interface prototype that integrates soft sensors and an inertial measurement unit. This study introduces both a rule-based data decoding method and a user intent inference-based decoding method to map human shoulder movements into continuous commands. Additionally, by incorporating prior knowledge of the user's operational performance into a shared autonomy framework, we implement an adaptive switching command mapping approach. This approach enables seamless transitions between the two decoding methods, enhancing their adaptability across different tasks.Main results.The proposed method has been validated on individuals with cervical spinal cord injury, bilateral arm amputation, and healthy subjects through a series of center-out target reaching tasks and a virtual powered wheelchair driving task. The experimental results show that using both the soft sensors and the gyroscope exhibits the most well-rounded performance in intent inference. Additionally, the rule-based method demonstrates better dynamic performance for wheelchair operation, while the intent inference method is more accurate but has higher latency. Adaptive switching decoding methods offer the best adaptability by seamlessly transitioning between decoding methods for different tasks. Furthermore, we discussed the differences and characteristics among the various types of participants in the experiment.Significance.The proposed method has the potential to be integrated into clothing, enabling non-invasive interaction with assistive devices in daily life, and could serve as a tool for rehabilitation assessment in the future.

Made with Artificial Intelligence: The Effect of Artificial Intelligence Disclosures in Instagram Advertisements on Consumer Attitudes

Artificial intelligence (AI) has the potential to disrupt the advertising industry as marketers and brands can leverage its power to create highly engaging personalized content. However, the usage of AI is prone to bias and misinformation and can be used to manipulate. Therefore, various lawmakers such as the European Union aim to enforce AI disclosure messages to protect consumers. But the implications of such disclosures have not yet been studied. This paper draws on existing theories in persuasion knowledge, disclosure theory, inferences of manipulative intent, and AI aversion to develop a model to understand consumer attitudes toward AI disclosures in Instagram advertisements. A three-condition between-subjects online experiment ( Nfinal = 161) was conducted to test the model. The data were analyzed using a moderated mediation model. AI disclosures lead to a direct decrease in advertising attitude. In addition, AI disclosures lead to a decrease in brand attitude only when consumers have high AI aversion. There were no effects of AI disclosures on source credibility. These effects were mediated by inferences of manipulative intent. However, participants who viewed the AI disclosure had lower inferences of manipulative intent and then participants who did not view the AI disclosure. Furthermore, no differences were found between AI disclosures pertaining to the use of AI in the creation of the image or the text. Implications are discussed from both theoretical and managerial viewpoints and highlight why the use of AI on social media for advertising purposes should be limited as it will become more transparent in the future.

Long-Term Tracking of Evasive Urban Target Based on Intention Inference and Deep Reinforcement Learning.

Unmanned aerial vehicles (UAVs) have been widely used in urban target-tracking tasks, where long-term tracking of evasive targets is of great significance for public safety. However, the tracked targets are easily lost due to the evasive behavior of the targets and the unstructured characteristics of the urban environment. To address this issue, this article proposes a hybrid target-tracking approach based on target intention inference and deep reinforcement learning (DRL). First, a target intention inference model based on convolution neural networks (CNNs) is built to infer target intentions by fusing urban environment information and observed target trajectory. Then, the prediction of the target trajectory can be inspired by the inferred target intentions, which can further provide effective guidance to the target search process. In order to fully explore the policy space, the target search policy is developed under a DRL framework, where the search policy is modeled as a deep neural network (DNN) and trained by interacting with the task environment. The simulation results show that the inference of the target intentions can effectively guide the UAV to search for the target and significantly improve the target-tracking performance. Meanwhile, the generalization results indicate that the proposed DRL-based search policy has high robustness to the uncertainty of the target behavior.

When Influencers Compare Brands: Defensive Processing of Sponsored Comparative Messages and Effects on Consumer Evaluations

Social media influencers comparing competing products helps draw the desirable consumer attention to their product recommendation. However, it may also risk backfiring, especially when the influencer post is sponsored by a given brand. In the present study, we explored consumer evaluations of sponsored influencer comparative marketing through two important defensive mechanisms—counterarguing and inferences of manipulative intent with a 2 (message appeal: noncomparative vs. comparative) × 2 (content sponsorship: organic vs. sponsored) between-subjects online experiment (N = 325). We found that when the post was sponsored, the comparative message generated significantly greater defensive reactions than the noncomparative message, leading to lower perceived trustworthiness and authenticity of the influencer, unfavorable brand attitude, and lower purchase intention. In contrast, when the post was organic, the observed difference diminished. Furthermore, we found more prominent effects via inferences of manipulative intent than counterarguing, suggesting that consumers considerably value the genuineness behind product promotion from influencers. These findings extend research on comparative advertising to the trending influencer marketing context, shedding novel light on mechanisms underlying consumer responses to sponsored comparative messaging.

Intention inference for space targets using deep convolutional neural network

Intention inference for space targets is crucial for space situational awareness. This paper introduces a rapid and precise method for recognizing the intentions of non-cooperative space targets using a deep convolutional neural network (CNN). By employing a relative orbital dynamics model, an analysis of relative motion was performed, resulting in the identification of 11 distinct motion intentions for space targets. This study also describes how to generate relative motion trajectory images to create a training set for intention inference, effectively converting the problem into one of image recognition and classification. Extensive simulations were carried out to fine-tune the network hyperparameters, and the results highlight the exceptional performance of the proposed CNN-based method, which achieved an accuracy of 99.682%. This method significantly enhances recognition accuracy over other neural network-based methods for space objects and offers considerable potential for applications like spacecraft collision avoidance and strategic maneuvers among space targets.

Not all communication styles are created equal: the differing effects of communicating the family business brand’s roots and virtues

PurposeThis study aims to examine the impact of communication style focusing on the family business brand’s roots and virtues on inferences of manipulative intent and willingness to pay a price premium, applying the persuasion knowledge model.Design/methodology/approachThis study collects data involving 337 participants from France and applies partial least squares structural equation modeling (PLS-SEM) to test the hypotheses.FindingsThe results demonstrate that communication style plays an important role in inferences of manipulative intent: focusing on the family business brand’s virtues has a positive effect while focusing on the family business brand’s roots has a nonsignificant effect. In turn, inferences of manipulative intent have a negative and significant effect on willingness to pay a price premium. Furthermore, age does not moderate the effect of communication style focusing on the family restaurant brand’s roots and inferences of manipulative intent but positively moderates the effect of communication style focusing on the family restaurant brand’s virtues on inferences of manipulative intent.Originality/valueTo the best of the authors’ knowledge, this study is the first to try to unpack the differing effects of communication styles in the context of family business brands. In this vein, it has insightful theoretical and managerial implications for family business brands.

Atypical and variable attention patterns reveal reduced contextual priors in children with autism spectrum disorder.

Accumulating evidence suggests that individuals with autism spectrum disorder (ASD) show impairments in using contextual priors to predict others' actions and make intention inference. Yet less is known about whether and how children with ASD acquire contextual priors during action observation and how contextual priors relate to their action prediction and intention inference. To form proper contextual priors, individuals need to observe the social scenes in a reliable manner and focus on socially relevant information. By employing a data-driven scan path method and areas of interest (AOI)-based analysis, the current study investigated how contextual priors would relate to action prediction and intention understanding in 4-to-9-year-old children with ASD (N = 56) and typically developing (TD) children (N = 50) during free viewing of dynamic social scenes with different intentions. Results showed that children with ASD exhibited higher intra-subject variability when scanning social scenes and reduced attention to socially relevant areas. Moreover, children with high-level action prediction and intention understanding showed lower intra-subject variability and increased attention to socially relevant areas. These findings suggest that altered fixation patterns might restrain children with ASD from acquiring proper contextual priors, which has cascading downstream effects on their action prediction and intention understanding.

Enhancing Robotic Collaborative Tasks Through Contextual Human Motion Prediction and Intention Inference

Enhancing Robotic Collaborative Tasks Through Contextual Human Motion Prediction and Intention Inference

Method of sequential intention inference for a space target based on meta-fuzzy decision tree

For an operational satellite in orbit, potential risks may arise when an unknown target approaches. Therefore, it is essential for the satellite to infer the intentions of the target in space. To achieve this objective, this paper proposes a method of sequential intention inference based on meta-fuzzy decision tree. The proposed method initially defines three situational functions: flyaround, collision, and observation. These functions help organize the inherent characteristics found in the trace and relative attitude of the target. Subsequently, by combining fuzzy decision tree (FDT), spatio-temporal modeling based on Long Short-Term Memory (LSTM), and task migration using meta-learning techniques, we can realize sequential intention inference for the target while ensuring quick adaptation to new tasks. Additionally, our proposed method takes into account the impact of fuzzy elements within FDT. Through quantum genetic algorithm (QGA) optimization, these fuzzy elements can be fine-tuned to enhance intention inference accuracy.

Peril Set-Aided Lane-Change Intention Inference on Highway

Peril Set-Aided Lane-Change Intention Inference on Highway

Hierarchical Context-Based Emotion Recognition With Scene Graphs.

For a better intention inference, we often try to figure out the emotional states of other people in social communications. Many studies on affective computing have been carried out to infer emotions through perceiving human states, i.e., facial expression and body posture. Such methods are skillful in a controlled environment. However, it often leads to misestimation due to the deficiency of effective inputs in unconstrained circumstances, that is, where context-aware emotion recognition appeared. We take inspiration from the advanced reasoning pattern of humans in perceived emotion recognition and propose the hierarchical context-based emotion recognition method with scene graphs. We propose to extract three contexts from the image, i.e., the entity context, the global context, and the scene context. The scene context contains abstract information about entity labels and their relationships. It is similar to the information processing of the human visual sensing mechanism. After that, these contexts are further fused to perform emotion recognition. We carried out a bunch of experiments on the widely used context-aware emotion datasets, i.e., CAER-S, EMOTIC, and BOdy Language Dataset (BoLD). We demonstrate that the hierarchical contexts can benefit emotion recognition by improving the accuracy of the SOTA score from 84.82% to 90.83% on CAER-S. The ablation experiments show that hierarchical contexts provide complementary information. Our method improves the F1 score of the SOTA result from 29.33% to 30.24% (C-F1) on EMOTIC. We also build the image-based emotion recognition task with BoLD-Img from BoLD and obtain a better emotion recognition score (ERS) score of 0.2153.

Multimodal fusion of EMG and vision for human grasp intent inference in prosthetic hand control.

Objective: For transradial amputees, robotic prosthetic hands promise to regain the capability to perform daily living activities. Current control methods based on physiological signals such as electromyography (EMG) are prone to yielding poor inference outcomes due to motion artifacts, muscle fatigue, and many more. Vision sensors are a major source of information about the environment state and can play a vital role in inferring feasible and intended gestures. However, visual evidence is also susceptible to its own artifacts, most often due to object occlusion, lighting changes, etc. Multimodal evidence fusion using physiological and vision sensor measurements is a natural approach due to the complementary strengths of these modalities. Methods: In this paper, we present a Bayesian evidence fusion framework for grasp intent inference using eye-view video, eye-gaze, and EMG from the forearm processed by neural network models. We analyze individual and fused performance as a function of time as the hand approaches the object to grasp it. For this purpose, we have also developed novel data processing and augmentation techniques to train neural network components. Results: Our results indicate that, on average, fusion improves the instantaneous upcoming grasp type classification accuracy while in the reaching phase by 13.66% and 14.8%, relative to EMG (81.64% non-fused) and visual evidence (80.5% non-fused) individually, resulting in an overall fusion accuracy of 95.3%. Conclusion: Our experimental data analyses demonstrate that EMG and visual evidence show complementary strengths, and as a consequence, fusion of multimodal evidence can outperform each individual evidence modality at any given time.

Manipulating Consumers with the Truth: Relative-Difference Claims in Advertising and Inferences of Manipulative Intent

Advertisements often feature numeric claims comparing risks associated with one alternative (e.g., control) to another (treatment). Differences may be expressed in either absolute (Starting Rate − Modified Rate) or relative ([Starting Rate − Modified Rate] / Starting Rate) terms. For example, assume a new heart disease treatment reduces the five-year mortality rate from 7.5% (starting risk) to 4% (modified risk). The absolute difference in this case is 3.5%, whereas the relative difference is 46.6%. Because relative differences appear larger, they positively influence evaluations and are therefore more often used in advertising. Unfortunately, baseline information (starting and modified rates) is often omitted or featured inconspicuously in an advertisement. In three experiments, we find that consumers ignore baseline information and instead rely on the magnitude of the difference in forming judgments (Study 1) unless they are motivated to elaborate on the relative-difference claim (Study 2). When motivated, inferences of manipulative intent (IMIs) are activated, which reduce evaluations. In Study 3, when motivated consumers’ ability to interpret a relative difference is enhanced, IMIs are affected by the magnitude of the absolute difference. Our findings suggest that motivated consumers or those with the ability to understand relative-difference claims find their use to be manipulative.

Coordination as inference in multi-agent reinforcement learning

The Centralized Training and Decentralized Execution (CTDE) paradigm, where a centralized critic is allowed to access global information during the training phase while maintaining the learned policies executed with only local information in a decentralized way, has achieved great progress in recent years. Despite the progress, CTDE may suffer from the issue of Centralized–Decentralized Mismatch (CDM): the suboptimality of one agent’s policy can exacerbate policy learning of other agents through the centralized joint critic. In contrast to centralized learning, the cooperative model that most closely resembles the way humans cooperate in nature is fully decentralized, i.e. Independent Learning (IL). However, there are still two issues that need to be addressed before agents coordinate through IL: (1) how agents are aware of the presence of other agents, and (2) how to coordinate with other agents to improve joint policy under IL. In this paper, we propose an inference-based coordinated MARL method: Deep Motor System (DMS). DMS first presents the idea of individual intention inference where agents are allowed to disentangle other agents from their environment. Secondly, causal inference was introduced to enhance coordination by reasoning each agent’s effect on others’ behavior. The proposed model was extensively experimented on a series of Multi-Agent MuJoCo and StarCraftII tasks. Results show that the proposed method outperforms independent learning algorithms and the coordination behavior among agents can be learned even without the CTDE paradigm compared to the state-of-the-art baselines including IPPO and HAPPO.

Obstacle Intention Awareness in Automatic Ship Collision Avoidance: Full-Scale Experiments in Confined Waters

In this article, full-scale experiments with a dynamic obstacle intention-aware Collision Avoidance System (CAS) are presented. The CAS consists of the Probabilistic Scenario-Based Model Predictive Control (PSB-MPC) for trajectory planning, dynamic obstacle avoidance, and antigrounding, with a Dynamic Bayesian Network (DBN) used for inferring obstacle intentions online. The novelty of this article lies in the utilization of intention information in deliberate collision-free planning. By inferring multiple different intention states on how and if nearby obstacles adhere to the COLREGS, the PSB-MPC can plan COLREGS-compliant avoidance maneuvers when possible, taking into account its awareness of the situation. The experiments put emphasis on hazardous situations where this intention information is both useful and necessary in order to avoid high collision risk. To the authors’ knowledge, the work is the first field experimental validation of such a probabilistic intention-aware CAS with consideration of multiple intention states. The experimental results demonstrate the validity of the proposed CAS scheme, with adherence to the traffic rules (COLREGS) 7, 8 and 13–17 in a diverse set of situations. The strengths and weaknesses of the proposed CAS are also discussed, giving insights that can be useful for researchers and practitioners in the field. Here, challenges related to detecting obstacle maneuvers and making the intention inference more robust to noise should be addressed as future work to make the scheme better suited for general usage on ships engaged in real traffic.

Adolescents with autism spectrum disorder exhibit intact physical causal inference but weak intention inference.

Individuals with autism spectrum disorder (ASD) show impaired mentalizing skills, specifically in understanding intentions. They have difficulty understanding social situations with multiple cues due to their limited ability to perceive subtle social contextual cues. Studies that used comic strips and the strange stories as intention attribution tests found that individuals with ASD exhibit a reduced ability in attributing intentions compared to inferring causal consequences. This study aims to use static photographs of social scenes taken in everyday settings to investigate the ability of adolescents with ASD to infer intentions in social contexts, and to explore how intention inference relates to working memory and basic attention, including sustained attention, selective attention, and divided attention. The results show that the physical causal inference ability of adolescents with ASD is comparable to typically developing adolescents, whereas intention inference is notably weaker. Furthermore, working memory predicts physical causal inference and divided attention predicts intention inference in ASD.

Bayesian Inference of Collision Avoidance Intent During Ship Encounters

Bayesian Inference of Collision Avoidance Intent During Ship Encounters

A Hierarchical Intention Inference Model for Connected and Automated Vehicles

A Hierarchical Intention Inference Model for Connected and Automated Vehicles