We investigate the approximation and estimation rates of conditional diffusion transformers (DiTs) with classifier-free guidance. We present a comprehensive analysis for “in-context” conditional DiTs under various common assumptions: generic and strong Hölder, linear latent (subspace), and Lipschitz score function assumptions. Importantly, we establish minimax optimality of DiTs by leveraging score function regularity. Specifically, we discretize the input domains into infinitesimal grids and then perform term-by-term Taylor expansions on the conditional diffusion score function under the Hölder smooth data assumption. This enables fine-grained use of transformers’ universal approximation through a more detailed piecewise constant approximation, and hence obtains tighter bounds. Additionally, we extend our analysis to latent settings. Our findings establish statistical limits for DiTs and offer practical guidance toward more efficient and accurate designs.

Flow matching is an emerging generative modeling framework that learns continuous-time dynamics to map noise into data. To enhance expressiveness and sampling efficiency, recent works have explored incorporating high-order trajectory information. Despite the empirical success, a holistic theoretical foundation is still lacking. We present a unified framework for standard and high-order flow matching that incorporates trajectory derivatives up to an arbitrary order $K$. Our key innovation is establishing the marginalization technique that converts the intractable $K$-order loss into a simple conditional regression with exact gradients and identifying the consistency constraint. We establish sharp statistical rates of the $K$-order flow matching implemented with transformer networks. With $n$ samples, flow matching estimates nonparametric distributions at a rate $\tilde{O}(n^{-\Theta(1/d )})$, matching minimax lower bounds up to logarithmic factors.

An Interval Type-2 (IT-2) fuzzy control approach is proposed to solve the Formation-Containment (F-C) problem for nonlinear multiple Autonomous Ship (AS) systems with the consideration of uncertainties and multiplicative noises. Based on the IT-2 Takagi-Sugeno Fuzzy Model (T-SFM), the nonlinear control problem can be recast into a linear problem. Different from the existing F-C control research, the formation problem is solved by the individual IT-2 fuzzy tracking controller of each leader. As a result, the information communication between leaders, who are farthest from each other, is needless. Moreover, other valuable functions such as the avoidance of known obstacles and time-varying formation can also be accomplished by designing the proper target trajectories for the tracking of leaders. By virtue of the IT-2 T-SFM, the analysis methods for linear Multi-Agent Control Systems (MACSs) can be utilized to solve the containment analysis problem, which is caused by the leaders' control input. Finally, the simulation results of multiple ASs are provided to demonstrate the advantage of the IT-2 fuzzy controller design approach in the F-C problem.

A Fuzzy Sliding Mode (FSM) tracking approach is proposed via Interval Type-2 (IT-2) Takagi-Sugeno Fuzzy Model (T-SFM) to complete Formation and Containment (F-and-C) tasks for Nonlinear Multi-Agent Systems (NMASs) with uncertainties and disturbances. Using Imperfect Premise Matching (IPM) approach, the IT-2 F-and-C Fuzzy Controllers (FCs) are designed in accordance with the IT-2 T-SFM. Different from the existing research, the individual tracking FC for each leader is proposed to complete the formation tasks and assignment of whole system dynamics. The stability analysis process is also simplified for ensuring the formation purpose. Benefiting from the advantage, the FSM control is conveniently combined into the FC design to upgrade the performance of leaders. For followers, the analysis methods of linear Multi-Agent (M-A) systems are extended to solve the containment analysis problem without the additional assumption by virtue of the IT-2 T-SFM representation. To deal with the disturbance effect, the passive constraints are combined into the analysis process. Finally, the proposed F-and-C FCs design method is applied to a Nonlinear Multi-Ships System (NMSS) to illustrate the advantage.

Identifying and characterizing the interaction between risk factors for multiple outcomes (multi-outcome interaction) has been one of the greatest challenges faced by complex multifactorial diseases. However, the existing approaches have several limitations in identifying the multi-outcome interaction. To address this issue, we proposed a multi-outcome interaction identification approach called MOAI. MOAI was motivated by the limitations of estimating the interaction simultaneously occurring in multi-outcomes and by the success of Pareto set filter operator for identifying multi-outcome interaction. MOAI permits the identification for the interaction of multiple outcomes and is applicable in population-based study designs. Our experimental results exhibited that the existing approaches are not effectively used to identify the multi-outcome interaction, whereas MOAI obviously exhibited superior performance in identifying multi-outcome interaction. We applied MOAI to identify the interaction between risk factors for colorectal cancer (CRC) in both metastases and mortality prognostic outcomes. An interaction between vaspin and carcinoembryonic antigen (CEA) was found, and the interaction indicated that patients with CRC characterized by higher vaspin (≥30%) and CEA (≥5) levels could simultaneously increase both metastases and mortality risk. The immunostaining evidence revealed that determined multi-outcome interaction could effectively identify the difference between non-metastases/survived and metastases/deceased patients, which offers multi-prognostic outcome risk estimation for CRC. To our knowledge, this is the first report of a multi-outcome interaction associated with a complex multifactorial disease. MOAI is freely available at https://sites.google.com/view/moaitool/home.

This study explores how the quality and credibility of information on healthcare websites can be enhanced through the simultaneous delivery of multiple information sources. Such information integration is achieved using a new experimental Web API called Portals. Accessing multiple information sources is salient due to the uncertainty surrounding an overwhelming amount of online health information that can be incorrect or misleading. When readers seek health information, the almost instantaneous verification that comes with the possibility of multiple-sources assessment is critical. This research provides novel insights that establish the value of reliably integrating health content from multiple sources (websites). The behavioral differences when people encounter consistent or inconsistent information from multiple-websites integration are investigated, and the implication of the findings are discussed. The research findings guide how health-related websites can help online seekers access higher-quality healthcare websites.

We present the Skin-Stroke Display, a system mounted on the lens inside the head-mounted display, which exerts subtle yet recognizable tactile feedback on the eye-ring using a motorized air jet. To inform our design of noticeable air-jet haptic feedback, we conducted a user study to identify absolute detection thresholds. Our results show that the tactile sensation had different sensitivity around the eyes, and we determined a standard intensity (8 mbar) to prevent turbulent airflow blowing into the eyes. In the second study, we asked participants to adjust the intensity around the eye for equal sensation based on standard intensity. Next, we investigated the recognition of point and stroke stimuli with or without inducing cognitive load on eight directions on the eye-ring. Our longStroke stimulus can achieve an accuracy of 82.6% without cognitive load and 80.6% with cognitive load simulated by the Stroop test. Finally, we demonstrate example applications using the skin-stroke display as the off-screen indicator, tactile I/O progress display, and tactile display.

Prior studies have evaluated various designs for audio notifications. However, calls for more in-depth research on how such notifications work, especially at the level of users' cognitive states, have gone unanswered; and studies evaluating audio notifications with large numbers of participants in multiple environments have been rare. This study conducted an electroencephalography study (N=20) and an online study (N=967) to enhance understandings of how three musical parameters - melody (simple, complex), pitch (high, low), and tempo (fast, slow) - influenced users' cognition and behaviors. There are eight different notifications with different combinations of these parameters. The online study analyzed the effects of user-specific and environmental information on users' behaviors while they listened to these notifications. The results revealed that tempo and pitch have the main effect on the speed and strength (accuracy) of users' cognition and behaviors. The users' characteristics and environments influenced the effects of these musical parameters.

Auditory alarms that repeatedly interrupt users until they react are common, especially in the context of alarms. However, when an alarm repeats, our brains habituate to it and perceive it less and less, with reductions in both perception and attention-shifting: a phenomenon known as the repetition-suppression effect (RS). To retain users' perception and attention, this paper proposes and tests the use of pitch- and intensity-modulated alarms. Its experimental findings suggest that the proposed modulated alarms can reduce RS, albeit in different patterns, depending on whether pitch or intensity is the focus of the modulation. Specifically, pitch-modulated alarms were found to reduce RS more when the number of repetitions was small, while intensity-modulated alarms reduced it more as the number of repetitions increased. Based on these results, we make several recommendations for the design of improved repeating alarms, based on which modulation approach should be adopted in various situations.

Graphic icons play an increasingly important role in interface design due to the proliferation of digital devices in recent years. Their ability to express information in a universal fashion allows us to immediately interact with new applications, systems, and devices. Icons can, however, cause user confusion and frustration if designed poorly. Several studies have evaluated icons using behavioral-performance metrics such as reaction time as well as self-report methods. However, determining the usability of icons based on behavioral measures alone is not straightforward, because users' interpretations of the meaning of icons involve various cognitive processes and perceptual mechanisms. Moreover, these perceptual mechanisms are affected not only by the icons themselves, but by usage scenarios. Thus, we need a means of sensitively and continuously measuring users' different cognitive processes when they are interacting with icons. In this study, we propose an EEG-based approach to icon evaluation, in which users' EEG signals are measured in multiple usage scenarios. Based on a combination of EEG and behavioral results, we provide a novel interpretation of the participants' perception during these tasks, and identify some important implications for icon design.