Triplet Adaptation Framework for Robust Semi-Supervised Learning.

Semi-supervised learning (SSL) suffers from severe performance degradation when labeled and unlabeled data come from inconsistent and imbalanced distribution. Nonetheless, there is a lack of theoretical guidance regarding a remedy for this issue. To bridge the gap between theoretical insights and practical solutions, we embark to an analysis of generalization bound of classic SSL algorithms. This analysis reveals that distribution inconsistency between unlabeled and labeled data can cause a significant generalization error bound. Motivated by this theoretical insight, we present a Triplet Adaptation Framework (TAF) to reduce the distribution divergence and improve the generalization of SSL models. TAF comprises three adapters: Balanced Residual Adapter, aiming to map the class distribution of labeled and unlabeled data to a uniform distribution for reducing class distribution divergence; Representation Adapter, aiming to map the representation distribution of unlabeled data to labeled one for reducing representation distribution divergence; and Pseudo-Label Adapter, aiming to align the predicted pseudo-labels with the class distribution of unlabeled data, thereby preventing erroneous pseudo-labels from exacerbating representation divergence. These three adapters collaborate synergistically to reduce the generalization bound, ultimately achieving a more robust and generalizable SSL model. Extensive experiments across various robust SSL scenarios validate the efficacy of our method.

How Reliable are Single-Question Workplace-Based Assessments in Surgery?

OBJECTIVE: Workplace-based assessments (WBAs) play an important role in the assessment of surgical trainees. Because these assessment tools are utilized by a multitude of faculty, inter-rater reliability is important to consider when interpreting WBA data. Although there is evidence supporting the validity of many of these tools, inter-reliability evidence is lacking. This study aimed to evaluate the inter-rater reliability of multiple operative WBA tools utilized in general surgery residency. DESIGN: General surgery residents and teaching faculty were recorded during 6 general surgery operations. Nine faculty raters each reviewed 6 videos and rated each resident on performance (using the Society for Improving Medical Professional Learning, or SIMPL, Performance Scale as well as the operative performance rating system (OPRS) Scale), entrustment (using the ten Cate Entrustment-Supervision Scale), and autonomy (using the Zwisch Scale). The ratings were reviewed for inter-rater reliability using percent agreement and intraclass correlations. PARTICIPANTS: Nine faculty members viewed the videos and assigned ratings for multiple WBAs. RESULTS: Absolute intraclass correlation coefficients for each scale ranged from 0.33 to 0.47. CONCLUSIONS: All single-item WBA scales had low to moderate inter-rater reliability. While rater training may improve inter-rater reliability for single observations, many observations by many raters are needed to reliably assess trainee performance in the workplace.

Enhancing Face Recognition With Detachable Self-Supervised Bypass Networks.

Attributed to the development of deep networks and abundant data, automatic face recognition (FR) has quickly reached human-level capacity in the past few years. However, the FR problem is not perfectly solved in case of large poses and uncontrolled occlusions. In this paper, we propose a novel bypass enhanced representation learning (BERL) method to improve face recognition under unconstrained scenarios. The proposed method integrates self-supervised learning and supervised learning together by attaching two auxiliary bypasses, a 3D reconstruction bypass and a blind inpainting bypass, to assist robust feature learning for face recognition. Among them, the 3D reconstruction bypass enforces the face recognition network to encode pose independent 3D facial information, which enhances the robustness to various poses. The blind inpainting bypass enforces the face recognition network to capture more facial context information for face inpainting, which enhances the robustness to occlusions. The whole framework is trained in end-to-end manner with two self-supervised tasks above and the classic supervised face identification task. During inference, the two auxiliary bypasses can be detached from the face recognition network, avoiding any additional computational overhead. Extensive experimental results on various face recognition benchmarks show that, without any cost of extra annotations and computations, our method outperforms state-of-the-art methods. Moreover, the learnt representations can also well generalize to other face-related downstream tasks such as the facial attribute recognition with limited labeled data.

Event Graph Guided Compositional Spatial-Temporal Reasoning for Video Question Answering.

Video question answering (VideoQA) is challenging since it requires the model to extract and combine multi-level visual concepts from local objects to global actions from complex events for compositional reasoning. Existing works represent the video with fixed-duration clip features that make the model struggle in capturing the crucial concepts in multiple granularities. To overcome this shortcoming, we propose to represent the video with an Event Graph in a hierarchical structure whose nodes correspond to visual concepts of different levels (object, relation, scene and action) and edges indicate their spatial-temporal relationships. We further propose a H ierarchical S patial- T emporal T ransformer (HSTT) which takes nodes from the graph as visual input to realize compositional reasoning guided by the event graph. To fully exploit the spatial-temporal context delivered from the graph structure, on the one hand, we encode the nodes in the order of their semantic hierarchy (depth) and occurrence time (breadth) with our improved graph search algorithm; On the other hand, we introduce edge-guided attention to combine the spatial-temporal context among nodes according to their edge connections. HSTT then performs QA by cross-modal interactions guaranteed by the hierarchical correspondence between the multi-level event graph and the cross-level question. Experiments on the recent challenging AGQA and STAR datasets show that the proposed method clearly outperforms the existing VideoQA models by a large margin, including those pre-trained with large-scale external data. Our code is available at https://github.com/ByZ0e/HSTT.

SurgNet: Self-Supervised Pretraining With Semantic Consistency for Vessel and Instrument Segmentation in Surgical Images.

Blood vessel and surgical instrument segmentation is a fundamental technique for robot-assisted surgical navigation. Despite the significant progress in natural image segmentation, surgical image-based vessel and instrument segmentation are rarely studied. In this work, we propose a novel self-supervised pretraining method (SurgNet) that can effectively learn representative vessel and instrument features from unlabeled surgical images. As a result, it allows for precise and efficient segmentation of vessels and instruments with only a small amount of labeled data. Specifically, we first construct a region adjacency graph (RAG) based on local semantic consistency in unlabeled surgical images and use it as a self-supervision signal for pseudo-mask segmentation. We then use the pseudo-mask to perform guided masked image modeling (GMIM) to learn representations that integrate structural information of intraoperative objectives more effectively. Our pretrained model, paired with various segmentation methods, can be applied to perform vessel and instrument segmentation accurately using limited labeled data for fine-tuning. We build an Intraoperative Vessel and Instrument Segmentation (IVIS) dataset, comprised of ~3 million unlabeled images and over 4,000 labeled images with manual vessel and instrument annotations to evaluate the effectiveness of our self-supervised pretraining method. We also evaluated the generalizability of our method to similar tasks using two public datasets. The results demonstrate that our approach outperforms the current state-of-the-art (SOTA) self-supervised representation learning methods in various surgical image segmentation tasks.

Variation in Competence of Graduating General Surgery Trainees.

OBJECTIVE: To examine the readiness of general surgery residents in their final year of training to perform 5 common surgical procedures based on their documented performance during training. DESIGN: Intraoperative performance ratings were analyzed using a Bayesian mixed effects approach, adjusting for rater, trainee, procedure, case complexity, and postgraduate year (PGY) as random effects as well as month in academic year and cumulative, procedure-specific performance per trainee as fixed effects. This model was then used to estimate each PGY 5 trainee's final probability of being able to independently perform each procedure. The actual, documented competency rates for individual trainees were then identified across each of the 5 most common general surgery procedures: appendectomy, cholecystectomy, ventral hernia repair, groin hernia repair, and partial colectomy. SETTING: This study was conducted using data from members of the SIMPL collaborative. PARTICIPANTS: A total of 17,248 evaluations of 927 PGY5 general surgery residents were analyzed from 2015 to 2021. RESULTS: The percentage of residents who requested a SIMPL rating during their PGY5 year and achieved a ≥90% probability of being rated as independent, or "Practice-Ready," was 97.4% for appendectomy, 82.4% for cholecystectomy, 43.5% for ventral hernia repair, 24% for groin hernia repair, and 5.3% for partial colectomy. CONCLUSIONS: There is substantial variation in the demonstrated competency of general surgery residents to perform several common surgical procedures at the end of their training. This variation in readiness calls for careful study of how surgical residents can become more adequately prepared to enter independent practice.

Mortality and Severe Complications Among Newly Graduated Surgeons in the United States.

OBJECTIVE: To evaluate severe complications and mortality over years of independent practice among general surgeons. BACKGROUND: Despite concerns that newly graduated general surgeons may be unprepared for independent practice, it is unclear whether patient outcomes differ between early and later career surgeons. METHODS: We used Medicare claims for patients discharged between July 1, 2007 and December 31, 2019 to evaluate 30-day severe complications and mortality for 26 operations defined as core procedures by the American Board of Surgery. Generalized additive mixed models were used to assess the association between surgeon years in practice and 30-day outcomes while adjusting for differences in patient, hospital, and surgeon characteristics. RESULTS: The cohort included 1,329,358 operations performed by 14,399 surgeons. In generalized mixed models, the relative risk (RR) of mortality was higher among surgeons in their first year of practice compared with surgeons in their 15th year of practice [5.5% (95% CI: 4.1%-7.3%) vs 4.7% (95% CI: 3.5%-6.3%), RR: 1.17 (95% CI: 1.11-1.22)]. Similarly, the RR of severe complications was higher among surgeons in their first year of practice compared with surgeons in their 15th year of practice [7.5% (95% CI: 6.6%-8.5%) versus 6.9% (95% CI: 6.1%-7.9%), RR: 1.08 (95% CI: 1.03-1.14)]. When stratified by individual operation, 21 operations had a significantly higher RR of mortality and all 26 operations had a significantly higher RR of severe complications in the first compared with the 15th year of practice. CONCLUSIONS: Among general surgeons performing common operations, rates of mortality and severe complications were higher among newly graduated surgeons compared with later career surgeons.

Evaluating Educational Outcomes Using Patient Outcomes of New Surgeons Performing Partial Colectomy Compared to Cholecystectomy.

PURPOSE: Despite ongoing efforts to improve surgical education, surgical residents face gaps in their training. However, it is unknown if differences in the training of surgeons are reflected in the patient outcomes of those surgeons once they enter practice. This study aimed to compare the patient outcomes among new surgeons performing partial colectomy-a common procedure for which training is limited-and cholecystectomy-a common procedure for which training is robust. METHOD: The authors retrospectively analyzed all adult Medicare claims data for patients undergoing inpatient partial colectomy and inpatient cholecystectomy between 2007 and 2018. Generalized additive mixed models were used to investigate the associations between surgeon years in practice and risk-adjusted rates of 30-day serious complications and death for patients undergoing partial colectomy and cholecystectomy. RESULTS: A total of 14,449 surgeons at 4,011 hospitals performed 340,114 partial colectomy and 355,923 cholecystectomy inpatient operations during the study period. Patients undergoing a partial colectomy by a surgeon in their 1st vs 15th year of practice had higher rates of serious complications (5.22% [95% CI, 4.85%-5.60%] vs 4.37% [95% CI, 4.22%-4.52%]; P < .01) and death (3.05% [95% CI, 2.92%-3.17%] vs 2.83% [95% CI, 2.75%-2.91%]; P < .01). Patients undergoing a cholecystectomy by a surgeon in their 1st vs 15th year of practice had similar rates of 30-day serious complications (4.11% vs 3.89%; P = .11) and death (1.71% vs 1.70%; P = .93). CONCLUSIONS: Patients undergoing partial colectomy faced a higher risk of serious complications and death when the operation was performed by a new surgeon compared to an experienced surgeon. Conversely, patient outcomes following cholecystectomy were similar for new and experienced surgeons. More attention to partial colectomy during residency training may benefit patients.

Electroacupuncture Exerts Chondroprotective Effect in Knee Osteoarthritis of Rabbits Through the Mitophagy Pathway.

Purpose: Mitochondrial dysfunction of chondrocytes has become an area of focus in Knee Osteoarthritis (KOA) in recent years. Activation of mitophagy could promote the survival of chondrocytes and alleviate cartilage degeneration. The aim of this study was to explore whether mitophagy was involved in the cartilage protection of KOA rabbits after electroacupuncture (EA) intervention. Methods: The rabbits were divided into 3 groups, Control group, KOA group, EA group, with 6 rabbits in each group. KOA model rabbits were established by modified Videman's extended immobilization method for 6 weeks and randomly divided into KOA group and EA group. The rabbits in EA group were treated every other day for 3 weeks. The degree of cartilage degeneration was detected by Safranine O-Fast Green staining and immunofluorescence. The morphological changes of chondrocytes mitochondria were detected by transmission electron microscope. ATP concentration in cartilage was measured by ATP Assay Kit. The changes of Pink1-Parkin signal pathway were detected by immunofluorescence, Western blot, and Real-time PCR. Results: The morphology showed that EA could reduce the degeneration of KOA cartilage and increase the distribution of collagen II. We also found that EA could activate mitophagy in KOA rabbit chondrocytes to remove damaged mitochondria and restore mitochondrial homeostasis, which was manifested as increasing the expression of LC3 II/I, promoting the colocalization of TOM20 and LC3B, reducing the accumulation of mitochondrial markers outer mitochondrial membrane 20 (TOM20) and inner mitochondrial membrane 23 (TIM23), and increasing ATP production in chondrocytes. This regulation might be achieved by upregulating the Pink1-Parkin signal pathway. Conclusion: EA may play a role in protecting KOA cartilage by activating mitophagy mediated through Pink1-Parkin pathway.

BLPSeg: Balance the Label Preference in Scribble-Supervised Semantic Segmentation.

Scribble-supervised semantic segmentation is an appealing weakly supervised technique with low labeling cost. Existing approaches mainly consider diffusing the labeled region of scribble by low-level feature similarity to narrow the supervision gap between scribble labels and mask labels. In this study, we observe an annotation bias between scribble and object mask, i.e., label workers tend to scribble on the spacious region instead of corners. This label preference makes the model learn well on those frequently labeled regions but poor on rarely labeled pixels. Therefore, we propose BLPSeg to balance the label preference for complete segmentation. Specifically, the BLPSeg first predicts an annotation probability map to evaluate the rarity of labels on each image, then utilizes a novel BLP loss to balance the model training by up-weighting those rare annotations. Additionally, to further alleviate the impact of label preference, we design a local aggregation module (LAM) to propagate supervision from labeled to unlabeled regions in gradient backpropagation. We conduct extensive experiments to illustrate the effectiveness of our BLPSeg. Our single-stage method even outperforms other advanced multi-stage methods and achieves state-of-the-art performance.

[Effect of acupotomy on expressions of cellular senescence markers p16Ink4a and p21Waf1/Cip1 in cartilage tissue of rabbits with knee osteoarthritis].

OBJECTIVE: To observe the effect of acupotomy on the expressions of p16Ink4a and p21Waf1/Cip1 in knee osteoarthritis (KOA) rabbits，so as to analyze whether acupotomy can treat KOA by inhibiting the cellular senescence of chondrocytes. METHODS: Twenty-four New Zealand male rabbits were randomly divided into normal, model, acupotomy and electroacupuncture (EA) groups, with 6 rabbits in each group. The KOA model was established by left hindlimb straightening fixation. After modeling, rabbits in the acupotomy group were treated with acupotomy loosening therapy on high stress points around the affected knee joints such as tendons attachment points of vastus medialis, vastus lateralis, rectus femoris, biceps femoris and pes anserine bursa, once a week for 3 weeks. In the EA group, "Xuehai"(SP10), "Liangqiu" (ST34),"Neixiyan" (EX-LE4) and "Waixiyan" (ST35) on the affected hindlimb were selected for EA treatment (3 mA, 2 Hz/100 Hz), 20 min each time, once every other day for 3 weeks. Before and after treatments, the knee Lequesne MG score and passive range of motion (PROM) of the affected knee joint were evaluated. After the treatments, the expressions of p16Ink4a and p21Waf1/Cip1 in the cartilage tissue of the affected knee joint were detected by immunohistochemistry and Western blot respectively. RESULTS: Before and after treatment, compared with the normal group, the Lequesne MG score was significantly increased (P<0.01), the PROM was significantly decreased (P<0.01) in the model group. After treatment, compared with the normal group, the positive expression and protein expression levels of p16Ink4a and p21Waf1/Cip1 were significantly increased (P<0.01) in the model group; compared with the model group, the Lequesne MG score was significantly decreased (P<0.01), the PROM was significantly increased (P<0.01), the positive expression and protein expression levels of p16Ink4a and p21Waf1/Cip1 were significantly decreased (P<0.01，P<0.05) in the acupo-tomy and EA groups; compared with the EA group, the Lequesne MG score was decreased (P<0.05), the PROM was increased (P<0.05), the positive expression and protein expression levels of p16Ink4a and p21Waf1/Cip1 were decreased (P<0.05，P<0.01) in the acupotomy group. CONCLUSION: Acupotomy intervention can down-regulate the expressions of cellular senescence markers p16Ink4a and p21Waf1/Cip1 in chondrocytes, indicating that acupotomy therapy may alleviate cartilage degeneration by inhibiting chondrocyte premature cellular senescence to treat KOA.

Early findings and strategies for successful implementation of SIMPL workplace-based assessments within vascular surgery residency and fellowship programs.

OBJECTIVE: As medical education systems increasingly move toward competency-based training, it is important to understand the tools available to assess competency and how these tools are utilized. The Society for Improving Medical Professional Learning (SIMPL) offers a smart phone-based assessment system that supports workplace-based assessment of residents' and fellows' operative autonomy, performance, and case complexity. The purpose of this study was to characterize implementation of the SIMPL app within vascular surgery integrated residency (0+5) and fellowship (5+2) training programs. METHODS: SIMPL operative ratings recorded between 2018 and 2022 were collected from all participating vascular surgery training institutions (n = 9 institutions with 5+2 and 0+5 programs; n = 4 institutions with 5+2 program only). The characteristics of programs, trainees, faculty, and SIMPL operative assessments were evaluated using descriptive statistics. RESULTS: Operative assessments were completed for 2457 cases by 85 attendings and 86 trainees, totaling 4615 unique operative assessment ratings. Attendings included dictated feedback in 52% of assessments. Senior-level residents received more assessments than junior-level residents (postgraduate year [PGY]1-3, n = 439; PGY4-5, n = 551). Performance ratings demonstrated increases from junior to senior trainees for both resident and fellow cohorts with "performance-ready" or "exceptional performance" ratings increasing by nearly two-fold for PGY1 to PGY5 residents (28.1% vs 40.6%), and from first- to second-year fellows (PGY6, 46.7%; PGY7, 60.3%). Similar gains in autonomy were demonstrated as trainees progressed through training. Senior residents were more frequently granted autonomy with "supervision only" than junior residents (PGY1, 8.7%; PGY5, 21.6%). "Supervision only" autonomy ratings were granted to 21.8% of graduating fellows. Assessment data included a greater proportion of complex cases for senior compared with junior fellows (PGY6, 20.9% vs PGY7, 26.5%). Program Directors felt that faculty and trainee buy-in were the main barriers to implementation of the SIMPL assessment app. CONCLUSIONS: This is the first description of the SIMPL app as an operative assessment tool within vascular surgery that has been successfully implemented in both residency and fellowship programs. The assessment data demonstrates expected progressive gains in trainees' autonomy and performance, as well as increasing case complexity, across PGY years. Given the selection of SIMPL as the assessment platform for required American Board of Surgery and Vascular Surgery Board Entrustable Professional Activities assessments, understanding facilitators and barriers to implementation of workplace-based assessments using this app is imperative, particularly as we move toward competency-based medical education.

Dual Compensation Residual Networks for Class Imbalanced Learning.

Learning generalizable representation and classifier for class-imbalanced data is challenging for data-driven deep models. Most studies attempt to re-balance the data distribution, which is prone to overfitting on tail classes and underfitting on head classes. In this work, we propose Dual Compensation Residual Networks to better fit both tail and head classes. First, we propose dual Feature Compensation Module (FCM) and Logit Compensation Module (LCM) to alleviate the overfitting issue. The design of these two modules is based on the observation: an important factor causing overfitting is that there is severe feature drift between training and test data on tail classes. In details, the test features of a tail category tend to drift towards feature cloud of multiple similar head categories. So FCM estimates a multi-mode feature drift direction for each tail category and compensate for it. Furthermore, LCM translates the deterministic feature drift vector estimated by FCM along intra-class variations, so as to cover a larger effective compensation space, thereby better fitting the test features. Second, we propose a Residual Balanced Multi-Proxies Classifier (RBMC) to alleviate the under-fitting issue. Motivated by the observation that re-balancing strategy hinders the classifier from learning sufficient head knowledge and eventually causes underfitting, RBMC utilizes uniform learning with a residual path to facilitate classifier learning. Comprehensive experiments on Long-tailed and Class-Incremental benchmarks validate the efficacy of our method.

Acupotomy Improves Synovial Hypoxia, Synovitis and Angiogenesis in KOA Rabbits.

Purpose: Knee osteoarthritis (KOA) is a chronic inflammatory disease highly associated with intra-articular hypertension, hypoxia and angiogenesis of synovial tissue. Our previous studies showed that acupotomy could treat KOA in a variety of ways, including reducing cartilage deterioration and enhancing biomechanical qualities. However, the mechanism of hypoxia and angiogenesis induced by acupotomy in KOA synovium remains unclear. This study looked for the benign intervention of acupotomy in synovial pathology. Methods: The rabbits were divided into 3 groups, Normal group, KOA group, and KOA + Acupotomy (Apo) group, with 11 rabbits in each group. The KOA rabbit model was established by the modified Videman method with six weeks. The KOA + Apo group performed the intervention. The tendon insertion of vastus medialis, vastus lateralis, rectus femoris, biceps femoris, and anserine bursa were selected as treatment points in rabbits. Rabbits were treated once every 7 days for 3 weeks. We observed the intra-articular pressure and oxygen partial pressure (BOLD MRI). The synovial morphology was monitored by Hematoxylin-Eosin Staining (HE Staining). The expression of hypoxia-inducible transcription factor-1α (HIF-1α), vascular endothelial growth factor (VEGF), interleukin-1ß (IL-1ß) and tumour necrosis factor-α (TNF-α) was detected using Immunohistochemical (IHC), Western Blot and Enzyme-Linked Immunosorbent Assay (ELISA). Results: Acupotomy reduced intra-articular hypertension and improved the synovial oxygen situation, synovial inflammatory and angiogenesis. HIF-1α, VEGF, IL-1ß and TNF-α expression were downregulated by acupotomy. Conclusion: Acupotomy may reduce inflammation and angiogenesis in KOA rabbit by reducing abnormally elevated intra-articular pressure and improving synovial oxygen environment. The above may provide a new theoretical foundation for acupotomy treatment of KOA.

Association of Surgical Resident Competency Ratings With Patient Outcomes.

PURPOSE: Accurate assessment of clinical performance is essential to ensure graduating residents are competent for unsupervised practice. The Accreditation Council for Graduate Medical Education milestones framework is the most widely used competency-based framework in the United States. However, the relationship between residents' milestones competency ratings and their subsequent early career clinical outcomes has not been established. It is important to examine the association between milestones competency ratings of U.S. general surgical residents and those surgeons' patient outcomes in early career practice. METHOD: A retrospective, cross-sectional study was conducted using a sample of national Medicare claims for 23 common, high-risk inpatient general surgical procedures performed between July 1, 2015, and November 30, 2018 (n = 12,400 cases) by nonfellowship-trained U.S. general surgeons. Milestone ratings collected during those surgeons' last year of residency (n = 701 residents) were compared with their risk-adjusted rates of mortality, any complication, or severe complication within 30 days of index operation during their first 2 years of practice. RESULTS: There were no associations between mean milestone competency ratings of graduating general surgery residents and their subsequent early career patient outcomes, including any complication (23% proficient vs 22% not yet proficient; relative risk [RR], 0.97, [95% CI, 0.88-1.08]); severe complication (9% vs 9%, respectively; RR, 1.01, [95% CI, 0.86-1.19]); and mortality (5% vs 5%; RR, 1.07, [95% CI, 0.88-1.30]). Secondary analyses yielded no associations between patient outcomes and milestone ratings specific to technical performance, or between patient outcomes and composites of operative performance, professionalism, or leadership milestones ratings ( P ranged .32-.97). CONCLUSIONS: Milestone ratings of graduating general surgery residents were not associated with the patient outcomes of those surgeons when they performed common, higher-risk procedures in a Medicare population. Efforts to improve how milestones ratings are generated might strengthen their association with early career outcomes.

Learning-Based Intrinsic Reflectional Symmetry Detection.

Reflectional symmetry is a ubiquitous pattern in nature. Previous works usually solve this problem by voting or sampling, suffering from high computational cost and randomness. In this article, we propose a learning-based approach to intrinsic reflectional symmetry detection. Instead of directly finding symmetric point pairs, we parametrize this self-isometry using a functional map matrix, which can be easily computed given the signs of Laplacian eigenfunctions under the symmetric mapping. Therefore, we manually label the eigenfunction signs for a variety of shapes and train a novel neural network to predict the sign of each eigenfunction under symmetry. Our network aims at learning the global property of functions and consequently converts the problem defined on the manifold to the functional domain. By disentangling the prediction of the matrix into separated bases, our method generalizes well to new shapes and is invariant under perturbation of eigenfunctions. Through extensive experiments, we demonstrate the robustness of our method in challenging cases, including different topology and incomplete shapes with holes. By avoiding random sampling, our learning-based algorithm is over 20 times faster than state-of-the-art methods, and meanwhile, is more robust, achieving higher correspondence accuracy in commonly used metrics.

CRIC: A VQA Dataset for Compositional Reasoning on Vision and Commonsense.

Alternatively inferring on the visual facts and commonsense is fundamental for an advanced visual question answering (VQA) system. This ability requires models to go beyond the literal understanding of commonsense. The system should not just treat objects as the entrance to query background knowledge, but fully ground commonsense to the visual world and imagine the possible relationships between objects, e.g., "fork, can lift, food". To comprehensively evaluate such abilities, we propose a VQA benchmark, Compositional Reasoning on vIsion and Commonsense(CRIC), which introduces new types of questions about CRIC, and an evaluation metric integrating the correctness of answering and commonsense grounding. To collect such questions and rich additional annotations to support the metric, we also propose an automatic algorithm to generate question samples from the scene graph associated with the images and the relevant knowledge graph. We further analyze several representative types of VQA models on the CRIC dataset. Experimental results show that grounding the commonsense to the image region and joint reasoning on vision and commonsense are still challenging for current approaches. The dataset is available at https://cricvqa.github.io.

Scaffold hopping via ring opening enables identification of acyclic compounds as new complement Factor D inhibitors.

The three complement pathways comprising the early phase of the complement system (the classical, lectin, and alternative pathways) act together with the innate and adaptive immune systems to protect against foreign entities and maintain tissue homeostasis. While these systems are normally under tight regulatory control, several diseases have been reported to correlate with uncontrolled activation and amplification of the alternative pathway, including paroxysmal nocturnal hemoglobinuria, atypical hemolytic uremic syndrome, C3 glomerulopathy, and age-related macular degeneration. Complement FactorD (CFD), a serine protease, is the rate-limiting enzyme for the activity of alternative pathway. CFD activates the alternative pathway by cleaving Complement Factor B complexed to C3b (C3bB) to generate alternative pathway C3 convertase (C3bBb). In our search for novel CFD inhibitors with therapeutic potential, we employed a hot-spot analysis of an ensemble of apo and holo CFD structures. This analysis identified potential pharmacophore features that aided in the design of a series of compounds based on an l-proline core. While these compounds inhibited CFD in an esterolytic assay (for example, a proline-based compound, IC50 = 161 nM), the pharmacokinetic (PK) properties were poor. A strategy of scaffold hopping via ring opening led to a novel series of acyclic compounds, with subsequent structure-based ligand design and lead optimization producing several novel CFD inhibitors. One of these inhibitors, 1-(2-((2-(3-chloro-2-fluorobenzylamino)-2-oxoethyl)(cyclopropyl)amino)-2-oxoethyl)-5-(3-methyl-3-(1-methylpiperidin-4-yl)ureido)-1H-indazole-3-carboxamide, showed good potency with IC50s of 37 nM in the esterolytic assay and 30 nM in a hemolytic assay and PK assessments following oral administration to rats revealed a Cmax of 113 ng/mL and an AUC0-24h of 257 hr.ng/mL.