Universal Adapter Research Articles

A functional trait-based assessment of urban street tree selection for Ethiopia

Urbanization in Ethiopia is rapidly reducing green space and biodiversity, and placing ecological stress on trees. Urban street trees (UST) can provide multiple ecosystem services, but need to cope with challenging street ecology and future climate risks. However, UST selection in Ethiopia is poor, and conventional trials and modern breeding are costly and time consuming to apply in Ethiopia. The main objective of the study was to identify potential UST species for two Ethiopian cities; Addis Ababa, and Arba Minch, by studying interspecific differences in functional traits. For the purpose, we created a database with list of 120 candidate UST gathered from multi-stakeholder workshop, literature review, and senior experts survey, and their ecological with their phenological and morphological features. Then we filtered 25 potential UST through framework analysis considering eco-morphological features of the trees. For the 25 species, we measured three types of plant functional traits: conservative traits (leaf mass per area and leaf dry matter content), acquisitive traits (leaf area, specific leaf area and leaf water content) and drought tolerance traits (leaf turgor pressure loss point, and leaf succulence index). Our results showed a strong trade-off between conservative and acquisitive traits, particularly between leaf dry matter content (LDMC) and specific leaf area (SLA) (r = −0.71), consistent with plant economic spectrum theory. Using PCA analysis, we identified three groups of potential USTs (Type I, II and III) that exhibit different adaptation strategies, consistent with Grime’s trait-based classification of universal plant adaptation strategies (CSR). Compared to species with high acquisitive traits (Type II and Type III), Type I species have high conservative and drought-tolerant traits shows better adaptation to harsh road environments, whereas Type II and Type III species may be more advantageous for urban parks and other resource-rich components of urban ecology. We also found significant variation in conservative, acquisitive and drought-tolerant traits among different species, indicating their different adaptation strategies. Our research advances the knowledge of plant adaptation in urban environments and provides a useful method for UST selection.

Reducing semantic ambiguity in domain adaptive semantic segmentation via probabilistic prototypical pixel contrast

Domain adaptation aims to reduce the model degradation on the target domain caused by the domain shift between the source and target domains. Although encouraging performance has been achieved by combining contrastive learning with the self-training paradigm, they suffer from ambiguous scenarios caused by scale, illumination, or overlapping when deploying deterministic embedding. To address these issues, we propose probabilistic prototypical pixel contrast (PPPC), a universal adaptation framework that models each pixel embedding as a probability via multivariate Gaussian distribution to fully exploit the uncertainty within them, eventually improving the representation quality of the model. In addition, we derive prototypes from probability estimation posterior probability estimation which helps to push the decision boundary away from the ambiguity points. Moreover, we employ an efficient method to compute similarity between distributions, eliminating the need for sampling and reparameterization, thereby significantly reducing computational overhead. Further, we dynamically select the ambiguous crops at the image level to enlarge the number of boundary points involved in contrastive learning, which benefits the establishment of precise distributions for each category. Extensive experimentation demonstrates that PPPC not only helps to address ambiguity at the pixel level, yielding discriminative representations but also achieves significant improvements in both synthetic-to-real and day-to-night adaptation tasks. It surpasses the previous state-of-the-art (SOTA) by +5.2% mIoU in the most challenging daytime-to-nighttime adaptation scenario, exhibiting stronger generalization on other unseen datasets. The code and models are available at https://github.com/DarlingInTheSV/Probabilistic-Prototypical-Pixel-Contrast.

A Universal Adaptive Algorithm for Graph Anomaly Detection

Graph-based anomaly detection aims to identify anomalous vertices in graph-structured data. It relies on the ability of graph neural networks (GNNs) to capture both relational and attribute information within graphs. However, previous GNN-based methods exhibit two critical shortcomings. Firstly, GNN is inherently a low-pass filter that tends to lead similar representations of neighboring vertices, which may result in the loss of critical anomalous information, termed as low-frequency constraints. Secondly, anomalous vertices that deliberately mimic normal vertices in features and structures are hard to detect, especially when the distribution of labels is unbalanced. To address these defects, we propose a Universal Adaptive Algorithm for Graph Anomaly Detection (U-A2GAD), which employs enhanced high frequency filters to overcome the low-frequency constraints, as well as aggregating both k-nearest neighbor (kNN) and k-farthest neighbor (kFN) to resolve the vertices’ camouflage problem. Extensive experiments demonstrated the effectiveness and universality of our proposed U-A2GAD and its constituent components, achieving improvements of up to 6% and an average increase of 2% on AUC-PR over the state-of-the-art methods. The source codes, and parameter setting details can be found at https://github.com/LIyvqi/U-A2GAD.

USDAP: universal source-free domain adaptation based on prompt learning

USDAP: universal source-free domain adaptation based on prompt learning

Robust Sequence Design Space for the Isothermal Exponential Amplification of Short Oligonucleotides.

Advances in isothermal amplification techniques have accelerated development in biosensing applications and the design of complex molecular devices. The exponential amplification reaction technique, or EXPAR, is uniquely positioned to process molecular information from short oligonucleotide strands (≈10nucleotides length) typically encountered in molecular computing or microRNA detection. Despite its conceptual simplicity (requiring only a template strand and two enzymes), the issue of nonspecific background amplification has hindered broader adoption. In this work, a new system configuration is established at 37 °C to achieve significantly improved performance. Critical sequence motifs responsible for the excellent signal-to-background profile are identified and generalized as a universal adapter design framework. Orthogonal template sequences generated from the framework are implemented for a triplex reaction and successfully evaluated mixtures of multiple-target inputs in a single-step, one-pot format without the need for exogenous agents.

Mix and Match Use of Revision Universal Head-Neck Adapters in Hip Arthroplasty: A Complications and Survival Analysis of 306 Cases

BackgroundOutcomes and safety of “mix and match” in total hip arthroplasty (THA) using universal head-neck adapters (UHNA) are a matter of ongoing discussion and concern due to legal affairs. This study aimed at analyzing the “mix and match” use of UHNA and evaluating complication and reoperation rates, possible risk factors, and the implant’s survival. MethodsA total of 306 patients treated with THA (94.1% revisions) using a UHNA at our institution between 2006 and 2022 were identified and included. Diagnoses, comorbidities, implants, and UHNA specifications were retrospectively recorded. Outcomes, complications, and survival analyses were evaluated, taking into account various possible risk factors. ResultsThere were 19.9% of the 306 included cases (58.5% women; median age 74 years; median follow-up 57 months) that had at least 1 complication. There were 43 patients (14.1%) who had to receive ≥1 rerevision surgery. The most common complication was postoperative recurrent dislocation (n = 27, 8.8%). There was one case of a prosthetic stem-neck fracture that was registered. Statistically significant risk factors for postoperative recurrent dislocations and postoperative aseptic loosening were, respectively, dislocation as an indication for UHNA implantation (P < .001) and oversized neck lengths (≥2XL; P = .004). The overall revision-free survival was 92% after 1 year and 82% at ten years. Statistically significant better survival rates were registered in patients ≥60 years old, who had fewer comorbidities (<2), and normal neck lengths (S to XL). ConclusionsThe results of this study underline the overall safety of UHNA use in THA through “mix and match.” Only one case of a stem-neck fracture was identified. The highlighted risk factors for failure must be kept in mind during the decision-making process with patients.

Development of 3D-printed universal adapter in enhancing retinal imaging accessibility

BackgroundThe revolutionary technology of smartphone-based retinal imaging has been consistently improving over the years. Smartphone-based retinal image acquisition devices are designed to be portable, easy to use, and cost-efficient, which enables eye care to be more widely accessible especially in geographically remote areas. This enables early disease detection for those who are in low- and middle- income population or just in general has very limited access to eye care. This study investigates the limitation of smartphone compatibility of existing smartphone-based retinal image acquisition devices. Additionally, this study aims to propose a universal adapter design that is usable with an existing smartphone-based retinal image acquisition device known as the PanOptic ophthalmoscope. This study also aims to simulate the reliability, validity, and performance overall of the developed prototype.MethodsA literature review has been conducted that identifies the limitation of smartphone compatibility among existing smartphone-based retinal image acquisition devices. Designing and modeling of proposed adapter were performed using the software AutoCAD 3D. For the proposed performance evaluation, finite element analysis (FEA) in the software Autodesk Inventor and 5-point scale method were demonstrated.ResultsPublished studies demonstrate that most of the existing smartphone-based retinal imaging devices have compatibility limited to specific older smartphone models. This highlights the benefit of a universal adapter in broadening the usability of existing smartphone-based retinal image acquisition devices. A functional universal adapter design has been developed that demonstrates its compatibility with a variety of smartphones regardless of the smartphone dimension or the position of the smartphone’s camera lens. The proposed performance evaluation method generates an efficient stress analysis of the proposed adapter design. The end-user survey results show a positive overall performance of the developed universal adapter. However, a significant difference between the expert's views on the developed adapter and the quality of images is observed.ConclusionThe compatibility of existing smartphone-based retinal imaging devices is still mostly limited to specific smartphone models. Besides this, the concept of a universal and suitable adapter for retinal imaging using the PanOptic ophthalmoscope was presented and validated in this paper. This work provides a platform for future development of smartphone-based ophthalmoscope that is universal.

Benchmarking a universal smoking machine adaptor for tobacco product testing

SignificanceCharacterisation of tobacco product emissions is an important step in assessing their impact on public health. Accurate and repeatable emissions data require that a leak-tight seal be made between the...

Automated combined unit with universal tillage adapter

The paper shows that using an automated combined unit with a universal tillage adapter for surface tillage represents a relevant and promising direction in the development of agricultural machinery. (Research purpose) The study aims to develop an automated combined unit equipped with a universal adapter designed to improve soil quality and increase the stability of the unit’s movement. (Materials and methods) The tillage adapter consists of two arc-shaped segments mounted at the front on a common axis that allows rotation. The rear sections contain a mini-hydraulic cylinder, which is remotely connected to the wireless actuator of the unit’s hydraulic system. The arcuate segments are equipped with replaceable cutting attachments and measure 25 centimeters in length. Mounted on a common axis with these segments, a holder is installed at the top at an angle of 40-45 degrees to the horizontal plane, while a saber-shaped slotter is positioned at the bottom at an angle of 15-20 degrees. (Results and discussion) It was discovered that the developed unit, equipped with an innovative adapter, facilitates three simultaneous operations: cutting the soil in a horizontal plane, loosening the soil, and eliminating weeds. Additionally, the working width of the device can be automated adjusted from the tractor cabin. The design of the unit improves tillage quality, increases movement stability, protects segments from wear and reduces manufacturing costs. (Conclusions) The proposed combined soil tillage unit provides crumbling, leveling, rolling and slicing of soil, and improves soil infi ltration. The location of the holder and slotting device with an inclination to the horizontal plane of 40-45 and 15-20 degrees, respectively, reduces the traction resistance of the unit and reduces energy costs.

Increasing Multienzyme Cascade Efficiency and Stability of MOF via Partitioning Immobilization.

Enhancing the stability of multienzyme cascade reactions in metal-organic frameworks (MOFs) is a challenging task in the fields of biotechnology and chemistry. However, addressing this challenge could yield far-reaching benefits across the application range in the biomedical, food, and environmental sectors. In this study, multienzyme partitioning immobilization that sequentially immobilizes cascade enzymes with hierarchical MOFs is proposed to reduce substrate diffusion resistance. Conversion results of ginsenosides indicate that this strategy improves the cascade efficiency up to 1.26 times. The substrate diffusion model is used to investigate the dual-interenzyme mass transfer behavior of substrates in the restricted domain space and evaluate the substrate channeling effect under partitioning immobilization. Molecular docking and kinetic simulations reveal that the MOFs effectively limit the conformational changes of cascade enzymes at high temperatures and in organic solvents while maintaining a large pocket of active centers. This phenomenon increased efficient substrate docking to the enzyme molecules, further optimizing cascade efficiency. The results of the immobilization of GOX and horseradish peroxidase as model enzymes indicate that the partitioned MOF immobilization strategy could be used for universal adaptation of cascade enzymes.

Curriculum adaptation method based on graph neural networks for universal domain adaptation

Universal domain adaptation (UniDA) aims to transfer knowledge between domains without prior knowledge of the label spaces. Category shift and domain shift are two primary challenges in UniDA, which require the method not only to distinguish common “known” samples from private “unknown” samples but also to align feature distributions of the common samples. In previous works, private samples have been involved to some extent in process of distribution alignment, leading to negative transfer. Therefore, we propose a curriculum adaptation method based on graph neural networks (GNN) to alleviate this issue. Specifically, our paper introduces a score to measure the transferability of samples and a curriculum adaptation strategy which only utilizes the most transferable samples for adaptation in each episode. To prevent classifier from overfitting on source samples, we assist classifier with the score and curriculum learning to obtain reliable pseudo-labels for target common samples. Furthermore, our method utilizes a graph neural network to aggregate features from similar samples across domains. Extensive experiments on three benchmarks demonstrate the effectiveness of our proposed framework.

Spa2 remodels ADP-actin via molecular condensation under glucose starvation.

Actin nucleotide-dependent actin remodeling is essential to orchestrate signal transduction and cell adaptation. Rapid energy starvation requires accurate and timely reorganization of the actin network. Despite distinct treadmilling mechanisms of ADP- and ATP-actin filaments, their filament structures are nearly identical. How other actin-binding proteins regulate ADP-actin filament assembly is unclear. Here, we show that Spa2 which is the polarisome scaffold protein specifically remodels ADP-actin upon energy starvation in budding yeast. Spa2 triggers ADP-actin monomer nucleation rapidly through a dimeric core of Spa2 (aa 281-535). Concurrently, the intrinsically disordered region (IDR, aa 1-281) guides Spa2 undergoing phase separation and wetting on the surface of ADP-G-actin-derived F-actin and bundles the filaments. Both ADP-actin-specific nucleation and bundling activities of Spa2 are actin D-loop dependent. The IDR and nucleation core of Spa2 are evolutionarily conserved by coexistence in the fungus kingdom, suggesting a universal adaptation mechanism in the fungal kingdom in response to glucose starvation, regulating ADP-G-actin and ADP-F-actin with high nucleotide homogeneity.

Maximum open-set entropy optimization via uncertainty measure for universal domain adaptation

Universal Domain Adaptation (UniDA) is a technology that enables the intelligent model to transfer knowledge learned from labeled source domains to related but unlabeled target domains without any prior label set relationship. The key to UniDA lies in rejecting target domain-specific “unknown” samples to achieve domain alignment on shared classes. In this paper, we propose the Maximum Open-set Entropy Optimization via uncertainty measure to adaptively reject “unknown” samples. Specifically, MOEO sets a transition region within the most easily confused class in the open-set space. We optimize the maximum open-set entropy of simple samples outside the transition region to further improve their confidence and separate difficult samples within the transition region by aggregating similar neighbors. Accordingly, a secure boundary is formed between shared samples and “unknown” samples, promoting domain alignment within shared classes. Experiments on four benchmarks show that MOEO outperforms the previous state-of-the-art significantly.

Development of a Universal Adaptive Control Algorithm for an Unknown MIMO System Using Recursive Least Squares and Parameter Self-Tuning

Development of a Universal Adaptive Control Algorithm for an Unknown MIMO System Using Recursive Least Squares and Parameter Self-Tuning

Enhanced Light-Matter Interaction in Metallic Nanoparticles: A Generic Strategy of Smart Void Filling.

The intrinsic properties of materials play a substantial role in light-matter interactions, impacting both bulk metals and nanostructures. While plasmonic nanostructures exhibit strong interactions with photons via plasmon resonances, achieving efficient light absorption/scattering in other transition metals remains a challenge, impeding various applications related to optoelectronics, chemistry, and energy harvesting. Here, we propose a universal strategy to enhance light-matter interaction, through introducing voids onto the surface of metallic nanoparticles. This strategy spans nine metals including those traditionally considered optically inactive. The absorption cross section of void-filled nanoparticles surpasses the value of plasmonic (Ag/Au) counterparts with tunable resonance peaks across a broad spectral range. Notably, this enhancement is achieved under arbitrary polarizations and varied particle sizes and in the presence of geometric disorder, highlighting the universal adaptability. Our strategy holds promise for inspiring emerging devices in photocatalysis, bioimaging, optical sensing, and beyond, particularly when metals other than gold or silver are preferred.

Bi-directional Adapter for Multimodal Tracking

Due to the rapid development of computer vision, single-modal (RGB) object tracking has made significant progress in recent years. Considering the limitation of single imaging sensor, multi-modal images (RGB, infrared, etc.) are introduced to compensate for this deficiency for all-weather object tracking in complex environments. However, as acquiring sufficient multi-modal tracking data is hard while the dominant modality changes with the open environment, most existing techniques fail to extract multi-modal complementary information dynamically, yielding unsatisfactory tracking performance. To handle this problem, we propose a novel multi-modal visual prompt tracking model based on a universal bi-directional adapter, cross-prompting multiple modalities mutually. Our model consists of a universal bi-directional adapter and multiple modality-specific transformer encoder branches with sharing parameters. The encoders extract features of each modality separately by using a frozen, pre-trained foundation model. We develop a simple but effective light feature adapter to transfer modality-specific information from one modality to another, performing visual feature prompt fusion in an adaptive manner. With adding fewer (0.32M) trainable parameters, our model achieves superior tracking performance in comparison with both the full fine-tuning methods and the prompt learning-based methods. Our code is available: https://github.com/SparkTempest/BAT.

An auto-regulated universal domain adaptation network for uncertain diagnostic scenarios of rotating machinery

In recent years, domain adaptation techniques have garnered significant attention in the field of intelligent fault diagnosis for mechanical equipment. Domain adaptation techniques can effectively enable the reuse of diagnostic knowledge, thereby improving the generalization performance of the trained model. However, many intelligent diagnostic models are tailored to specific diagnostic scenarios, including closed-set domain adaptation, partial domain adaptation, and open-set domain adaptation. When these models built for specific scenarios are applied to other scenarios, their diagnostic performance will deteriorate catastrophically. This study proposes an auto-regulated universal domain adaptation network (AUDAN) for uncertain diagnostic scenarios of rotating machinery, which can handle multiple diagnostic scenarios and adaptively set threshold boundaries for each fault category based on data distribution structure to distinguish known and unknown samples. Specifically, a novel integration loss function is designed to train the proposed AUDAN. The self-adaptive reweight module is introduced into the proposed method to distribute the relative weight for each loss function automatically. Experiments on two rotating machinery datasets are carried out for validations. The experimental results demonstrate that the proposed AUDAN performs effectively in uncertain diagnostic scenarios and is promising in addressing the fault classification tasks in real industries.

Abstract 6709: P329G-Engager: A novel universal antibody-based adaptor platform for cancer immunotherapy

Abstract Bispecific immune cell engagers are showing promise in cancer immunotherapy. To achieve optimal efficacy, there is a growing need for patient personalization, tackling intra-tumor heterogeneity, and simplified combination therapies. Here, we present a novel modular anti-P329G adaptor platform aiming to address these issues, which is based on the recognition of the P329G mutation in the Fc portion of a targeting adaptor antibody. This P329G mutation can serve to recruit different anti-P329G effector cell engagers, including innate cell engagers, T cell engagers, costimulators or immunocytokines. Specifically, upon identifying patient-specific tumor surface targets, the primary antigen-binding IgG1 antibodies bearing Fc-silencing P329G LALA mutations are applied. Subsequently, based on the tumor’s immune profile, the best-suited anti-tumoral effector cell type is chosen for recruitment and activation. Based on this information, the secondary antibody recognizing the P329G mutation is chosen from an array of effector cell engagers with different modes of action: P329G-T cell bispecifics (P329G-TCB), ADCC-competent P329G innate cell engagers (P329G-ICE), P329G costimulatory molecules (P329G-CD28/4-1BBL) or P329G immunocytokines (P329G-IL2v). In vitro assays showed all P329G modalities inducing anti-tumoral and/or immunomodulatory cell activity. Anti-tumoral efficacy of secondary antibodies was observed only in presence of tumor-targeted P329G adaptors, while no effect was observed in their absence. The readouts consisted of tumor killing quantification, CD4+ and CD8+ T cell activation, cytokine release and cytokine receptor signaling. As a proof of concept study in vivo, an experiment in MKN-45 (CEACAM5+)-bearing humanized mice was performed, using an anti-CEACAM5 P329G-IgG adaptor and a P329G-TCB. Both tumor volume shrinkage and T cell infiltration into the tumor confirmed anti-tumoral efficacy of the platform, as compared to a conventional CEACAM5-targeted TCB. Neither the individual P329G IgG nor the individual P329G-TCB induced anti-tumoral efficacy, validating the requirement for primary and secondary antibody binding for T cell engaging activity. These results provide in vitro and in vivo evidence that the universal P329G engager platform can be used as an efficacious cancer treatment. Ultimately, this modular approach may enable off-the-shelf personalization via combination of patient-specific antibodies and universal effector cell engagers based on the patient’s tumor target and immune profile. Citation Format: Marlena Surowka, Idil Hutter-Karakoc, Diana Darowski, Christina Claus, Claudia Ferrara Koller, Anne Freimoser-Grundschober, Thomas Hofer, Andrzej Sobieniecki, Denis Assisi, Stephane Leclair, Ekkehard Moessner, Pablo Umaña, Maria Amann, Christian Klein. P329G-Engager: A novel universal antibody-based adaptor platform for cancer immunotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6709.

A universal multi-source domain adaptation method with unsupervised clustering for mechanical fault diagnosis under incomplete data

Recently, due to the difficulty of collecting condition data covering all mechanical fault types in industrial scenarios, the fault diagnosis problem under incomplete data is receiving increasing attention where no target prior information can be available. The existing open-set or universal domain adaptation (DA) diagnosis methods typically treat private fault samples in the target as a generalized “unknown” fault class, neglecting their inherent structure. This oversight can lead to confusion in latent feature space representations and difficulties in separating unknown samples. Therefore, a universal DA method with unsupervised clustering is developed to explore the intrinsic structure of the target samples for mechanical fault diagnosis, where multi-source information on different working conditions is considered to transfer complementary knowledge. First, a composite clustering metric combining single-domain and cross-domain evaluation is constructed to recognize shared and unknown health classes on source–target domains. Second, to alleviate the intra-class shift while enlarging the inter-class gap, a class-wise DA algorithm is suggested which operates on the basis of maximum mean discrepancy. Finally, an entropy regularization criterion is utilized to facilitate clustering of different health classes. The efficacy of the presented approach in the fault diagnosis issues when monitoring data is inadequate has been verified through extensive experiments on three rotating machinery datasets.

Universal Adaptive Construction of Verifiable Secret Sharing and Its Application to Verifiable Secure Distributed Data Storage

Universal Adaptive Construction of Verifiable Secret Sharing and Its Application to Verifiable Secure Distributed Data Storage