About

Tao Gao is an Associate Professor jointly appointed to the Department of Communication and the Department of Statistics at UCLA. He explores human social perception and cognition, focusing on aspects of the human mind that can inspire the development of artificial intelligence that is communicative and trustworthy. His research aims to reverse engineer human intuitive social commonsense by examining how humans perceive social concepts such as animacy, intentions, beliefs, desires, emotions, personality, and morality from static images or short videos. Gao's work seeks to reveal the nature of human social commonsense through cognitive modeling and psychophysical experiments, with the goal of implementing human-like social understanding in artificial intelligence and robots to enable seamless, safe, and trustworthy communication and collaboration with humans. His interdisciplinary research draws tools from cognitive science, artificial intelligence, robotics, and computer vision. He obtained his Ph.D. in cognitive psychology from Yale University, completed a post-doctoral fellowship at MIT's Center of Brain, Mind and Machine, and previously worked as a research scientist in the Computer Vision and Machine Learning labs at GE Global Research.

Research topics

• Computer Science
• Artificial Intelligence
• Natural Language Processing
• Human–computer interaction
• Psychology
• Epistemology
• Cognitive science
• Control engineering
• Physics
• Structural engineering
• Mathematics
• Engineering

Selected publications

• Communicating Through Acting: Affording Communicative Intention in Pantomimes

  Cognitive Science · 2025-10-01

  articleSenior author

  How do people intuitively recognize communicative intention in pantomimes, even though such actions kinematically resemble instrumental behaviors directed at changing the world? We focus on two alternative hypotheses: one posits that instrumental intention competes with communicative intention, such that the weaker the former, the stronger the latter; the other suggests that instrumental intention is nested within communicative intention, such that the presence of the former facilitates the latter. To test these hypotheses, we compiled a video dataset of action-object pairs with varying frequencies in the English corpus. Using the concept of affordance, we qualitatively varied the degree to which a scene visually supports the execution of an action. Across two empirical experiments, we found a nonmonotonic relationship between affordance and communicative ratings: partial affordance, where the scene provides some support for an action's instrumental purpose, elicited the strongest perception of communicative intention. In contrast, full affordance or no affordance resulted in weaker interpretations of communicative intention. We also found that recognizing the instrumental components of pantomime-like actions predicted a higher communicativeness rating. Our study, on top of confirming humans' ability to interpret novel pantomimes, reveals a novel mechanism of communicative intention: recognizing an instrumental goal and perceiving suboptimal conditions for achieving it together enhance the communicative signal. This work contributes toward an integrated theory of pantomimes, demonstrating how the rationality principle not only aids in distinguishing communicative intention but also supports the identification of instrumental content embedded within it.

  PublisherDOI

• Model-Based Control Strategies Comparison of One Bionic Ankle Tensegrity Exoskeleton: BATE

  2025-05-19 · 2 citations

  article

  This paper presents a comparative analysis of model-based control strategies for a Bionic Ankle Tensegrity Exoskeleton (BATE), designed to emulate the self-stress equilibrium and self-supporting characteristics of the human ankle biotensegrity structure. Model-based control strategies are conventional methods that can discover the principles of the BATE exoskeleton. The high dimensions and non-linearity of the BATE pose challenges for theoretical modeling and model-based control strategies. To address this, we propose a modeling method based on the force density that accounts for interaction forces. We evaluated the trajectory tracking performance and robustness of BATE under three power-assisted control methods: position control (PC), force control (FC) and hybrid force-position control (FPC). Experimental results demonstrate that the PC method offers superior performance in both trajectory tracking and robustness, making it suitable for early rehabilitation training to enhance flexibility. Our findings highlight the advantages of tensegrity exoskeletons over current wearable exoskeletons and introduce novel concepts for developing high-performance exoskeletons.

  PublisherDOI

• Concealment and detection: The influence of management tone on analyst forecast revisions

  International Review of Financial Analysis · 2025-01-28 · 2 citations

  article
  PublisherDOI

• Observer-Based Fixed-Time Distributed Formation Tracking Control for Fixed-Wing UAVs Under External Disturbances

  2025-05-16

  article1st authorCorresponding

  This paper investigates the observer-based fixedtime formation tracking method of fixed-wing UAVs subject to unknown external disturbances. To realize the desired formation configuration, a hierarchical control strategy consisting of distributed observer and formation controller is proposed. Firstly, a distributed fixed-time observer for each follower is proposed to estimate the state information of the leader, which promises the fixed-time convergence. Meanwhile, a non-singular fixed-time sliding-mode tracking controller is proposed, which eliminates the influence of external disturbances and overcomes the singularity of the terminal sliding-mode control. Finally, the effectiveness of the proposed control strategy is verified by numerical simulation.

  PublisherDOI

• Amip-N: Ankle Motion Intention Prediction Network Combined Tcn, Lstm, and Wavelet Multi-Head Self-Attention Based on Imus

  SSRN Electronic Journal · 2025-01-01

  preprintOpen access
  PublisherDOI

• A Biomechanical Evaluation of a Novel Interspinous Process Device: In Vitro Flexibility Assessment and Finite Element Analysis

  Bioengineering · 2025-04-03 · 2 citations

  articleOpen access

  The interspinous process device (IPD) has emerged as a viable alternative for managing lumbar degenerative pathologies. Nevertheless, limited research exists regarding mechanical failure modes including device failure and spinous process fracture. This study developed a novel IPD (IPD-NEW) and systematically evaluated its biomechanical characteristics through finite element (FE) analysis and in vitro cadaveric biomechanical testing. Six human L1-L5 lumbar specimens were subjected to mechanical testing under four experimental conditions: (1) Intact spine (control); (2) L3-L4 implanted with IPD-NEW; (3) L3-L4 implanted with Wallis device; (4) L3-L4 implanted with Coflex device. Segmental range of motion (ROM) was quantified across all test conditions. A validated L1-L5 finite element model was subsequently employed to assess biomechanical responses under both static and vertical vibration loading regimes. Comparative analysis revealed that IPD-NEW demonstrated comparable segmental ROM to the Wallis device while exhibiting lower rigidity than the Coflex implant. The novel design effectively preserved physiological spinal mobility while enhancing load distribution capacity. IPD-NEW demonstrated notable reductions in facet joint forces, device stress concentrations, and spinous process loading compared to conventional implants, particularly under vibrational loading conditions. These findings suggest that IPD-NEW may mitigate risks associated with facetogenic pain, device failure, and spinous process fracture through optimized load redistribution.

  PublisherOA PDFDOI

• A relevance model of human sparse communication in cooperation

  Frontiers in Robotics and AI · 2025-07-30

  articleOpen accessSenior authorCorresponding

  Human real-time communication creates a limitation on the flow of information, which requires the transfer of carefully chosen and condensed data in various situations. We introduce a model that explains how humans choose information for communication by utilizing the concept of "relevance" derived from decision-making theory and Theory of Mind (ToM). We evaluated the model by conducting experiments where human participants and an artificial intelligence (AI) agent assist each other to avoid multiple traps in a simulated navigation task. The relevance model accurately depicts how humans choose which trap to communicate. It also outperforms GPT-4, which participates in the same task by responding to prompts that describe the game settings and rules. Furthermore, we demonstrated that when humans received assisting information from an AI agent, they achieved a much higher performance and gave higher ratings to the AI when it utilized the relevance model compared to a heuristic model. Together, these findings provide compelling evidence that a relevance model rooted in decision theory and ToM can effectively capture the sparse and spontaneous nature of human communication.

  PublisherOA PDFDOI

• Simulation Study of Parallel Permanent Magnet Electromagnetic Hybrid Suspension System

  Chinese Rare Earths · 2025-04-01

  articleOpen access1st authorCorresponding

  The paper proposes a parallel hybrid levitation structure for the uncontrollable levitation problem faced by the rail transportation mechanism of repulsive type levitation realized by rare earth permanent magnet array. In the ANSYS simulation environment, the mechanical characteristics of the repulsive levitation structure based on Halbach permanent magnet array are analyzed, and the nonlinear distribution characteristics of the static magnetic field also make the levitation mechanism have nonlinear mechanical relationship and large levitation drop, the vehicle attitude within the controllable range does not affect the permanent magnetic levitation capability. The dynamic performance adjustment effect of the mechanical damping structure and the electromagnetic damping structure on the permanent magnet suspension array is explored through systematic simulation classification. The active control ability of the electromagnetic damping system makes the permanent magnet suspension system have controllability and stability. The relationship between the electromagnetic damping system and the vehicle permanent magnet on the levitation guidance performance under the influence of spatial position factors is investigated, and the results show that the electromagnetic damping system cannot achieve the synergistic control of two degrees of freedom, but the additional electromagnetic damping constraint can effectively solve the control problem of permanent magnet levitation and improve the levitation stiffness to suppress the problems of large oscillation dropout and fast magnetic energy decay.

  PublisherOA PDFDOI

• The influence of emotional intelligence on career decision-making difficulties: a moderated chain mediating model

  Current Psychology · 2025-03-12

  article
  PublisherDOI

• Coordinated Control Strategy for Output Voltage of PEMFC System Based on Multi-Agent Deep Deterministic Policy Gradient

  2025-05-16

  article1st authorCorresponding

  In this paper, in order to improve the stability of the output voltage of proton exchange membrane fuel cell (PEMFC) power generation system, the coupling relationship between the oxygen flow rate and the duty cycle of the DC/DC converter is investigated, and the Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm is used to treat the two as independent agents, and to realize the coordinated control of the two under different loading conditions through parallel learning and adaptive coordination. Simulations show that the method effectively improves the output voltage stability under different load conditions with good robustness.

  PublisherDOI

Frequent coauthors

• Zhengguang Liu

  13 shared

• Jun Zhang

  Dalian Maritime University

  12 shared

• Guo Li

  Hebei University of Economics and Business

  11 shared

• Mowei Shen

  University of California, Berkeley

  10 shared

• Brian J. Scholl

  Yale University

  10 shared

• Juanjuan Zhang

  Fudan University

  9 shared

• S Jialiang

  Nanjing General Hospital of Nanjing Military Command

  9 shared

• L Jieshou

  Nanjing General Hospital of Nanjing Military Command

  9 shared