Xiang Li
Loading...
7 results
Filters
Reset filtersSearch Results
Publications1 - 7 of 7
- PVT v2: Improved baselines with Pyramid Vision TransformerItem type: Journal Article
Computational Visual MediaWang, Wenhai; Xie, Enze; Li, Xiang; et al. (2022)Transformers have recently lead to encouraging progress in computer vision. In this work, we present new baselines by improving the original Pyramid Vision Transformer (PVT v1) by adding three designs: (i) a linear complexity attention layer, (ii) an overlapping patch embedding, and (iii) a convolutional feed-forward network. With these modifications, PVT v2 reduces the computational complexity of PVT v1 to linearity and provides significant improvements on fundamental vision tasks such as classification, detection, and segmentation. In particular, PVT v2 achieves comparable or better performance than recent work such as the Swin transformer. We hope this work will facilitate state-of-the-art transformer research in computer vision. Code is available at https://github.com/whai362/PVT. - Achieving Near-Optimal Convergence for Distributed Minimax Optimization with Adaptive StepsizesItem type: Conference Paper
Advances in Neural Information Processing Systems 37Huang, Yan; Li, Xiang; Shen, Yipeng; et al. (2024)In this paper, we show that applying adaptive methods directly to distributed minimax problems can result in non-convergence due to inconsistency in locally computed adaptive stepsizes. To address this challenge, we propose D-AdaST, a Distributed Adaptive minimax method with Stepsize Tracking. The key strategy is to employ an adaptive stepsize tracking protocol involving the transmission of two extra (scalar) variables. This protocol ensures the consistency among stepsizes of nodes, eliminating the steady-state error due to the lack of coordination of stepsizes among nodes that commonly exists in vanilla distributed adaptive methods, and thus guarantees exact convergence. For nonconvex-strongly-concave distributed minimax problems, we characterize the specific transient times that ensure time-scale separation of stepsizes and quasi-independence of networks, leading to a near-optimal convergence rate of $\tilde{\mathcal{O}} \left( ε^{-\left( 4+δ\right)} \right)$ for any small $δ> 0$, matching that of the centralized counterpart. To our best knowledge, D-AdaST is the first distributed adaptive method achieving near-optimal convergence without knowing any problem-dependent parameters for nonconvex minimax problems. Extensive experiments are conducted to validate our theoretical results. - Biomechanical Evaluation of a Novel Loop Retention Mechanism for Cortical Graft Fixation in ACL ReconstructionItem type: Journal Article
Orthopaedic Journal of Sports MedicineGötschi, Tobias; Rosenberg, George; Li, Xiang; et al. (2020)Background: Implant fixation by means of a cortical fixation device (CFD) has become a routine procedure in anterior cruciate ligament reconstruction. There is no clear consensus whether adjustable-length CFDs are more susceptible to loop lengthening when compared with pretied fixed-length CFDs. Purpose: To assess biomechanical performance measures of 3 types of CFDs when subjected to various loading protocols. Study Design: Controlled laboratory study. Methods: Three types of CFDs underwent biomechanical testing: 1 fixed length and 2 adjustable length. One of the adjustable-length devices is based on the so-called finger trap mechanism, and the other is based on a modified sling lock mechanism. A device-only test of 5000 cycles (n = 8 per group) and a tendon-device test of 1000 cycles (n = 8 per group) with lower and upper force limits of 50 and 250 N, respectively, were applied, followed by ramp-to-failure testing. Adjustable-length devices then underwent further cyclic testing with complete loop unloading (n = 5 per group) at each cycle, as well as fatigue testing (n = 3 per group) over a total of 1 million cycles. Derived mechanical parameters were compared among the devices for statistical significance using Kruskal-Wallis analysis of variance followed by post hoc Mann-Whitney U testing with Bonferroni correction. Results: All CFDs showed elongation <2 mm after 5000 cycles when tested in an isolated manner and withstood ultimate tensile forces in excess of estimated peak in vivo forces. In both device-only and tendon-device tests, differences in cyclic performance were found among the devices, favoring adjustable-length fixation devices over the fixed-length device. Completely unloading the suspension loops, however, led to excessive loop lengthening of the finger trap device, whereas the modified sling lock device remained stable throughout the test. The fixed-length device displayed superior ultimate strength over both adjustable-length devices. Both adjustable-length devices showed adequate fatigue behavior during high-cyclic testing. Conclusion: All tested devices successfully prevented critical construct elongation when tested with constant tension and withstood ultimate loads in excess of estimated in vivo forces during the rehabilitation phase. The finger trap device gradually lengthened excessively when completely unloaded during cyclic testing. - Desulfurization of 2-phenylcyclohexanethiol over transition-metal phosphidesItem type: Journal Article
Journal of CatalysisZhou, Xuerong; Li, Xiang; Prins, Roel; et al. (2020) - Metal organic framework-assisted in-situ synthesis of β-NiMnOOH nanosheets with abundant NiOOH active sites for efficient electro-oxidation of ureaItem type: Journal Article
Journal of Colloid and Interface ScienceYan, Xiaodong; Li, Xiang; Zhang, Wen-Da; et al. (2023)NiOOH has been considered as the active center for urea oxidation reaction (UOR), but it remains challenging to synthesize high-performance NiOOH-based catalysts. Herein, we realize the synthesis of a high-performance NiOOH-based catalyst through in-situ transformation from the NiMn-based metal–organic framework to NiMnOOH. X-ray photoelectron spectroscopy characterization shows that the Ni3+/Ni2+ ratio in the NiMnOOH is 3.9 times as big as that in the Ni(OH)2, and in-situ Raman characterization further consolidates the presence of the NiOOH species in the NiMnOOH and as well unveils the faciliated Ni2+/Ni3+ redox reaction. The abundant NiOOH species, the markedly facilitated Ni2+/Ni3+ redox reaction and the Ni-Mn synergy contribute to the high intrinsic activity of the NiMnOOH towards UOR. The NiMnOOH exhibits an impressively low onset potential of 1.305 V vs reversible hydrogen electrode (RHE) and requires only a small potential of 1.34 V vs RHE to deliver a current density of 100 mA cm−2 in 1.0 M KOH + 0.33 M urea. In addition, the NiMnOOH catalyst possesses good long-term working stability. - Multimodal Remote Home Monitoring of Lung Transplant Recipients during COVID-19 Vaccinations: Usability Pilot Study of the COVIDA Desk Incorporating Wearable DevicesItem type: Journal Article
MedicinaSchuurmans, Macé M.; Muszynski, Michal; Li, Xiang; et al. (2023)Abstract Background and Objectives: Remote patient monitoring (RPM) of vital signs and symptoms for lung transplant recipients (LTRs) has become increasingly relevant in many situations. Nevertheless, RPM research integrating multisensory home monitoring in LTRs is scarce. We developed a novel multisensory home monitoring device and tested it in the context of COVID-19 vaccinations. We hypothesize that multisensory RPM and smartphone-based questionnaire feedback on signs and symptoms will be well accepted among LTRs. To assess the usability and acceptability of a remote monitoring system consisting of wearable devices, including home spirometry and a smartphone-based questionnaire application for symptom and vital sign monitoring using wearable devices, during the first and second SARS-CoV-2 vaccination. Materials and Methods: Observational usability pilot study for six weeks of home monitoring with the COVIDA Desk for LTRs. During the first week after the vaccination, intensive monitoring was performed by recording data on physical activity, spirometry, temperature, pulse oximetry and self-reported symptoms, signs and additional measurements. During the subsequent days, the number of monitoring assessments was reduced. LTRs reported on their perceptions of the usability of the monitoring device through a purpose-designed questionnaire. Results: Ten LTRs planning to receive the first COVID-19 vaccinations were recruited. For the intensive monitoring study phase, LTRs recorded symptoms, signs and additional measurements. The most frequent adverse events reported were local pain, fatigue, sleep disturbance and headache. The duration of these symptoms was 5–8 days post-vaccination. Adherence to the main monitoring devices was high. LTRs rated usability as high. The majority were willing to continue monitoring. Conclusions: The COVIDA Desk showed favorable technical performance and was well accepted by the LTRs during the vaccination phase of the pandemic. The feasibility of the RPM system deployment was proven by the rapid recruitment uptake, technical performance (i.e., low number of errors), favorable user experience questionnaires and detailed individual user feedback. - Transformation of the Swiss Energy System for a Net-Zero Greenhouse Gas Emission SocietyItem type: ReportPanos, Evangelos; Kober, Tom; Ramachandran, Kannan; et al. (2021)
Publications1 - 7 of 7