This work targets to merge various Vision Transformers (ViTs) trained on different tasks (i.e., datasets with different object categories) or domains (i.e., datasets with the same categories but different environments) into one unified model, yielding still good performance on each task or domain. Previous model merging works focus on either CNNs or NLP models, leaving the ViTs merging research untouched. To fill this gap, we first explore and find that existing model merging methods cannot well handle the merging of the whole ViT models and still have improvement space. To enable the merging of the whole ViT, we propose a simple-but-effective gating network that can both merge all kinds of layers (e.g., Embedding, Norm, Attention, and MLP) and select the suitable classifier. Specifically, the gating network is trained by unlabeled datasets from all the tasks (domains), and predicts the probability of which task (domain) the input belongs to for merging the models during inference. To further boost the performance of the merged model, especially when the difficulty of merging tasks increases, we design a novel metric of model weight similarity, and utilize it to realize controllable and combined weight merging. Comprehensive experiments on kinds of newly established benchmarks, validate the superiority of the proposed ViT merging framework for different tasks and domains. Our method can even merge beyond 10 ViT models from different vision tasks with a negligible effect on the performance of each task.
The recent advances in autonomous systems have prompted a strong demand for the next generation of adaptive structures and materials to possess more built-in intelligence in their mechanical domain, the so-called mechano-intelligence (MI). Previous MI attempts mainly focused on specific designs and case studies to realize limited aspects of MI, and there is a lack of a systematic foundation in constructing and integrating the different elements of intelligence in an effective and efficient manner. Here, we propose a new approach to create the needed foundation in realizing integrated multifunctional MI via a physical reservoir computing (PRC) framework. That is, to concurrently embody computing power and the various elements of intelligence, namely perception, decision-making, and commanding, directly in the mechanical domain, advancing from conventional adaptive structures that rely solely on add-on digital computers and massive electronics to achieve intelligence. As an exemplar platform, we construct a mechanically intelligent phononic metastructure with the integrated elements of MI by harnessing the PRC power hidden in their high-degree-of-freedom nonlinear dynamics. Through analyses and experimental investigations, we uncover multiple adaptive structural functions ranging from self-tuning wave controls to wave-based logic gates. This research will provide the basis for creating future new structures that would greatly surpass the state of the art - such as lower power consumption, more direct interactions, and much better survivability in harsh environment or under cyberattacks. Moreover, it will enable the addition of new functions and autonomy to systems without overburdening the onboard computers.
A framework is proposed for developing and evaluating algorithms for extracting multipath propagation components (MPCs) from measurements collected by sounders at millimeter-wave (mmW) frequencies. To focus on algorithmic performance, an idealized model is proposed for the spatial frequency response of the propagation environment measured by a sounder. The input to the sounder model is a pre-determined set of MPC parameters that serve as the "ground truth." A three-dimensional angle-delay (beamspace) representation of the measured spatial frequency response serves as a natural domain for implementing and analyzing MPC extraction algorithms. Metrics for quantifying the error in estimated MPC parameters are introduced. Initial results are presented for a greedy matching pursuit algorithm that performs a least-squares (LS) reconstruction of the MPC path gains within the iterations. The results indicate that the simple greedy-LS algorithm has the ability to extract MPCs over a large dynamic range, and suggest several avenues for further performance improvement through extensions of the greedy-LS algorithm as well as by incorporating features of other algorithms, such as SAGE and RIMAX.
One of the main benefits of a wrist-worn computer is its ability to collect a variety of physiological data in a minimally intrusive manner. Among these data, electrodermal activity (EDA) is readily collected and provides a window into a person's emotional and sympathetic responses. EDA data collected using a wearable wristband are easily influenced by motion artifacts (MAs) that may significantly distort the data and degrade the quality of analyses performed on the data if not identified and removed. Prior work has demonstrated that MAs can be successfully detected using supervised machine learning algorithms on a small data set collected in a lab setting. In this paper, we demonstrate that unsupervised learning algorithms perform competitively with supervised algorithms for detecting MAs on EDA data collected in both a lab-based setting and a real-world setting comprising about 23 hours of data. We also find, somewhat surprisingly, that incorporating accelerometer data as well as EDA improves detection accuracy only slightly for supervised algorithms and significantly degrades the accuracy of unsupervised algorithms.