Search

Yan Xie,1 Wenni Zhou,1 Xiang Tao,2 Hui Lv,3 Zhongping Cheng1 1Department of Obstetrics and Gynecology, Shanghai Tenth People’s Hospital, School of Medicine, Tongji University, Shanghai, 200072, People’s Republic of China; 2Department of Pathology, Obstetrics and Gynecology Hospital of Fudan University, Shanghai, 200090, People’s Republic of China; 3SG+ Bio-Testing Inc, Shanghai, 200093, People’s Republic of ChinaCorrespondence: Zhongping Cheng; Hui Lv, Email mdcheng18@tongji.edu.cn; lvhui@sgplus.cnObjective: Gestational diabetes mellitus (GDM) and preeclampsia (PE) are common pregnancy complications that share some common risk factors. GDM patients are also at high risk for PE. There are no sensitive markers for prediction, especially for the occurrence of PE in GDM patients. This study investigated plasma proteins for the prediction of PE in GDM patients.Methods: A total of 10 PE, 10 GDM, and 5 PE complicated with GDM cases, as well as 10 pregnant controls without obvious complications, were included in the nested cohort. The proteomics in the plasma collected at 12– 20 weeks of gestational age (GA) were analyzed by liquid chromatography‒mass spectrometry/mass spectrometry. Some potential markers, such as soluble transferrin receptor (sTfR), ceruloplasmin (CP), apolipoprotein E (ApoE) and inositol 1,4,5-trisphosphate receptor 1 (ITPR1), were validated using enzyme-linked immunosorbent assays.Results: Functional analysis of the plasma showed that proteasome activation, pancreatic secretion, and fatty acid degradation were activated in the GDM group, and renin secretion-, lysosome-, and proteasome pathways involving iron transport and lipid metabolism were enriched in the PE group, distinguishing PE complicating GDM.Conclusion: Through proteomics analysis of plasma in early pregnancy, PE complicating GDM may have a unique mechanism from that of PE alone. Plasma sTfR, CP and ApoE levels have potential clinical applications in early screening.Keywords: gestational diabetes mellitus, preeclampsia, preeclampsia complicating gestational diabetes mellitus, proteomics, plasma biomarkers

Blocking increased expression of nerve injury-specific long non-coding RNA (NIS-lncRNA) in injured dorsal root ganglia (DRG) through DRG microinjection of NIS-lncRNA small hairpin interfering RNA or generation of NIS-lncRNA knockdown mice mitigates neuropathic pain. However, these strategies are impractical in the clinic. This study employed a Food and Drug Administration (FDA)-approved antisense oligonucleotides strategy to examine the effect of NIS-lncRNA ASOs on neuropathic pain.Effects of intrathecal injection of NIS-lncRNA antisense oligonucleotides on day 7 or 14 after chronic constriction injury (CCI) of the sciatic nerve, fourth lumbar (L4) spinal nerve ligation, or intraperitoneal injection of paclitaxel or streptozotocin on the expression of DRG NIS-lncRNA and C-C chemokine ligand 2 (CCL2, an NIS-lncRNA downstream target) and nociceptive hypersensitivity were examined. We also assessed whether NIS-lncRNA antisense oligonucleotides produced cellular toxicity.Intrathecal NIS-lncRNA antisense oligonucleotides attenuated CCI-induced mechanical allodynia, heat hyperalgesia, cold hyperalgesia, and ongoing nociceptive responses, without changing basal or acute nociceptive responses and locomotor function. Intrathecal NIS-lncRNA antisense oligonucleotides also blocked CCI-induced increases in NIS-lncRNA and CCL2 in the ipsilateral L3 and L4 DRG and hyperactivities of neurones and astrocytes in the ipsilateral L3 and L4 spinal cord dorsal horn. Similar results were found in antisense oligonucleotides-treated mice after spinal nerve ligation or intraperitoneal injection of paclitaxel or streptozotocin. Normal morphologic structure and no cell loss were observed in the DRG and spinal cord of antisense oligonucleotides-treated mice.These findings further validate the role of NIS-lncRNA in trauma-, chemotherapy-, or diabetes-induced neuropathic pain and demonstrate potential clinical application of NIS-lncRNA antisense oligonucleotides for neuropathic pain management.

Dysregulation of pain-associated genes in the dorsal root ganglion (DRG) is considered to be a molecular basis of neuropathic pain genesis. Fused in sarcoma (FUS), a DNA/RNA-binding protein, is a critical regulator of gene expression. However, whether it contributes to neuropathic pain is unknown. This study showed that peripheral nerve injury caused by the fourth lumbar (L4) spinal nerve ligation (SNL) or chronic constriction injury (CCI) of the sciatic nerve produced a marked increase in the expression of FUS protein in injured DRG neurons. Blocking this increase through microinjection of the adeno-associated virus (AAV) 5-expressingFusshRNA into the ipsilateral L4 DRG mitigated the SNL-induced nociceptive hypersensitivities in both male and female mice. This microinjection also alleviated the SNL-induced increases in the levels of phosphorylated extracellular signal-regulated kinase 1/2 (p-ERK1/2) and glial fibrillary acidic protein (GFAP) in the ipsilateral L4 dorsal horn. Furthermore, mimicking this increase through microinjection of AAV5 expressing full-lengthFusmRNA into unilateral L3/4 DRGs produced the elevations in the levels of p-ERK1/2 and GFAP in the dorsal horn, enhanced responses to mechanical, heat and cold stimuli, and induced the spontaneous pain on the ipsilateral side of both male and female mice in the absence of SNL. Mechanistically, the increased FUS activated the NF-κB signaling pathway by promoting the translocation of p65 into the nucleus and phosphorylation of p65 in the nucleus from injured DRG neurons. Our results indicate that DRG FUS contributes to neuropathic pain likely through the activation of NF-κB in primary sensory neurons.SIGNIFICANCE STATEMENTIn the present study, we reported that fused in sarcoma (FUS), a DNA/RNA-binding protein, is upregulated in injured dorsal root ganglion (DRG) following peripheral nerve injury. This upregulation is responsible for nerve injury-induced translocation of p65 into the nucleus and phosphorylation of p65 in the nucleus from injured DRG neurons. Because blocking this upregulation alleviates nerve injury-induced nociceptive hypersensitivity, DRG FUS participates in neuropathic pain likely through the activation of NF-κB in primary sensory neurons. FUS may be a potential target for neuropathic pain management.

Uterine endometrial cancer (UEC) is an estrogen-related tumor. Succinate and heme metabolism play important roles in the progression of multiple tumors. However, the relationship between estrogen, succinate, and heme metabolism and related regulatory mechanisms remain largely unknown. In this study, we observed that the expression of aminolevulinate delta synthase 1 (ALAS1) and solute carrier family member 38 (SLC25A38) in UEC tissues is significantly higher than that in normal tissues. Further analysis showed that estrogen and succinate increased the expression of ALAS1 and SLC25A38 in uterine endometrial cancer cells (UECC), and the administration of succinate upregulated the level of the estrogen receptor (ER). Silencing nuclear receptor coactivator 1 (NCOA1) reversed the effects of estrogen and succinate via downregulation of ALAS1 expression. Additionally, exposure of UECC to heme increased cell viability and invasiveness, while silencing the NCOA1 gene weakened this effect. These findings revealed that estrogen and succinate can synergistically increase the expression of ALAS1 and SLC25A38 via the ERβ/NCOA1 axis, promoting heme accumulation and increasing the proliferative and invasive potential of UECC.

Breast cancer is one of the most frequent malignancies in females. The molecular mechanism of how breast cancer development and recurrence still need to be explored. Peroxisome gamma coactivator-1β (PGC-1β) was engaged in cancer energy metabolism and tumor genesis. However, the mechanisms of PGC-1β in breast cancer have not been fully understood. In this study, PCG-1β overexpressed and knockdown vectors were transferred into MCF-7 cells. With the association-quantitative connection analysis, the different expressions of mRNAs and proteins were examined. Additionally, the terms on differentially expressed mRNAs and proteins were enriched by GO and KEGG. Based on the results, 1872 differentially expressed genes were identified in the up-regulated of PGC-1β group, and 1318 genes were found in the down-regulated of PGC-1β cells. With the label-free technique, 221 differentially expressed proteins were screened in PGC-1β up-regulated group, and 459 proteins were identified in PGC-1β down-regulated group. Correlation analysis showed that 49 significantly expressed mRNA-protein pairs in OV vs CT groups and 25 paired in SI vs CT groups. Combined analysis of transcriptome and proteome demonstrated that PGC-1β plays a important role in cancer energy metabolism and boosting the pace of chemical processes in the proliferation of breast cancer cells. Additional investigation about PGC-1β and energy metabolism in cancer cells may shed fresh light on the growth and treatment of breast cancer cells.

PurposeThe stress–strain behaviors of rockfill materials in dams are significantly affected by the anisotropy and grain crushing. However, these factors are rarely considered in numerical simulations of high rockfill dams. This study intends to develop a reasonable and practical constitutive model for rockfill materials to overcome the above problems.Design/methodology/approachThe effects of anisotropy and grain crushing are comprehensively considered by the spatial position of the reference state line. After the improved generalized plasticity model for rockfill materials (referred to as the PZR model) is developed and verified by laboratory tests, it is used with the finite element method to simulate the stress–strain behaviors of the Nuozhadu high core rockfill dam.FindingsThe simulated results agree well with the laboratory tests data and the situ monitoring data, verifying the reliability and practicability of the developed PZR model.Originality/valueA new anisotropic state parameter is proposed to reflect the nonmonotonic variation in the strength as the major principal stress direction angle varies. This advantage is verified by the simulation of a set of conventional triaxial tests with different inclination angles of the compaction plane. 2) This is the first time that the elastoplastic model is verified by the situ monitoring data of high core rockfill dams. The numerical simulation results show that the PZR model can well reflect the stress–strain characteristics of rockfill materials in high core rockfill dams and is better than the traditional EB model.

Hypoxia-induced M2 phenotypes of tumor associated macrophages (TAMs) promote the development and chemoresistance of multiple types of cancers, including glioblastoma (GBM). However, the detailed molecular mechanisms have not been fully understood. In this study, we firstly reported that hypoxic pressure promoted M2 macrophage generation, which further promoted cancer progression and temozolomide (TMZ) resistance in GBM through secreting vascular endothelial growth factor (VEGF). Specifically, the clinical data suggested that M2 macrophages were significantly enriched in GBM tissues compared with the adjacent normal tissues, and the following in vitro experiments validated that hypoxic pressure promoted M2-polarized macrophages through upregulating hypoxia-inducible factor-1α (HIF-1α). In addition, hypoxic M2 macrophages VEGF-dependently promoted cell proliferation, epithelial-mesenchymal transition (EMT), glioblastoma stem cell (GSC) properties, and TMZ resistance in GBM cells through activating the PI3K/Akt/Nrf2 pathway. Also, M2 macrophages secreted VEGF to accelerate angiogenesis in human umbilical vein endothelial cells (HUVECs) through interacting with its receptor VEGFR. In general, we concluded that hypoxic M2 macrophages contributed to cancer progression, stemness, drug resistance, and angiogenesis in GBM through secreting VEGF, and our data supported the notion that targeting hypoxia-associated M2 macrophages might be an effective treatment strategy for GBM in clinical practices.

Necrotizing funisitis is a distinct lesion of the umbilical cord associated with chorioamnionitis and bloodborne fetal infection. The lesion may be a response to microorganisms in Wharton's jelly. A common microorganism detected in chorioamnionitis is Ureaplasma urealyticum (U. urealyticum). This study hypothesizes that U. urealyticum DNA will be present in Wharton's jelly in necrotizing funisitis.Necrotizing funisitis was identified retrospectively from a 2-year pathology database and confirmed in review. Paraffin fixed embedded tissue sections of the lesion were prepared for polymerase chain reaction (PCR) by using primers to identify U. urealyticum. Twenty matched controls without funisitis were similarly processed. Clinical data included serological tests of common bloodborne infections in the mothers and infants, and U. urealyticum PCR results in the urine of the neonates.Fourteen cases of necrotizing funisitis were identified in 7,416 examined placentas. Nine of these umbilical cords were positive by PCR for U. urealyticum (64.3%). Nineteen of twenty control cases were negative. Eight of ten neonates (80%) also had positive urine PCR tests for U. urealyticum. No infants or mothers had evidence of bloodborne fetal infection.U. urealyticum DNA was present in Wharton's jelly by PCR testing in the majority of the necrotizing funisitis lesions tested. This result supports a possible causative role for U. urealyticum in many cases of necrotizing funisitis.

Background. Ferroptosis is a nonapoptotic form of programmed cell death, which may be related to the occurrence and development of sepsis-induced acute respiratory distress syndrome (ARDS)/acute lung injury (ALI). Mucin 1 (MUC1) is a kind of macromolecule transmembrane glycoprotein. Previous studies have shown that MUC1 could relieve ALI in sepsis and predict whether sepsis patients would develop into ARDS. However, the role of MUC1 in the ferroptosis of sepsis-induced ALI/ARDS remains unclear. Materials and Methods. Sera samples from 50 patients with sepsis/septic shock were used to detect iron metabolism-related markers. Western blot and qRT-PCR were conducted to detect the expression levels of ferroptosis-related genes. Enzyme-linked immunosorbent assay (ELISA) was performed to evaluate inflammatory factors. Transmission electron microscopy (TEM) was used to assess morphological changes of cells. Results. The results showed that the iron metabolism-related indicators in sepsis-induced ARDS patients changed significantly, suggesting the iron metabolism disorder. The expression levels of ferroptosis-related genes in lung tissues of sepsis had marked changes, and the lipid peroxidation levels increased, while Ferrostatin-1 (Fer-1) could reverse the above results, which confirmed the occurrence of ferroptosis. In terms of mechanism studies, inhibition of MUC1 dimerization could increase the expression level of Keap1, reduce the phosphorylation level of GSK3β, inhibit the entry of Nrf2 into the nucleus, further inhibit the expression level of GPX4, enhance the lipid peroxidation level of lung tissues, trigger ferroptosis, and aggravate lung injury. Besides, inhibiting MUC1 reversed the alleviating effect of vitamin E on ALI caused by sepsis, increased the aggregation of inflammatory cells in lung tissues, and aggravated alveolar injury and edema. Conclusions. Our study was the first to explore the changes of iron metabolism indicators in ALI/ARDS of sepsis, clarify the important role of ferroptosis in ALI/ARDS induced by sepsis, and reveal the effects and specific mechanisms of MUC1 in regulating ferroptosis, as well as the sensitization on vitamin E.

3D shape modeling is labor-intensive and time-consuming and requires years of expertise. Recently, 2D sketches and text inputs were considered as conditional modalities to 3D shape generation networks to facilitate 3D shape modeling. However, text does not contain enough fine-grained information and is more suitable to describe a category or appearance rather than geometry, while 2D sketches are ambiguous, and depicting complex 3D shapes in 2D again requires extensive practice. Instead, we explore virtual reality sketches that are drawn directly in 3D. We assume that the sketches are created by novices, without any art training, and aim to reconstruct physically-plausible 3D shapes. Since such sketches are potentially ambiguous, we tackle the problem of the generation of multiple 3D shapes that follow the input sketch structure. Limited in the size of the training data, we carefully design our method, training the model step-by-step and leveraging multi-modal 3D shape representation. To guarantee the plausibility of generated 3D shapes we leverage the normalizing flow that models the distribution of the latent space of 3D shapes. To encourage the fidelity of the generated 3D models to an input sketch, we propose a dedicated loss that we deploy at different stages of the training process. We plan to make our code publicly available.

In contrast to the standard learning paradigm where all classes can be observed in training data, learning with augmented classes (LAC) tackles the problem where augmented classes unobserved in the training data may emerge in the test phase. Previous research showed that given unlabeled data, an unbiased risk estimator (URE) can be derived, which can be minimized for LAC with theoretical guarantees. However, this URE is only restricted to the specific type of one-versus-rest loss functions for multi-class classification, making it not flexible enough when the loss needs to be changed with the dataset in practice. In this paper, we propose a generalized URE that can be equipped with arbitrary loss functions while maintaining the theoretical guarantees, given unlabeled data for LAC. To alleviate the issue of negative empirical risk commonly encountered by previous studies, we further propose a novel risk-penalty regularization term. Experiments demonstrate the effectiveness of our proposed method., Accepted by AAAI 2023, 17 pages, 4 tables, 4 figures

Recently, the claim of "near-ambient superconductivity" in a N-doped lutetium hydride attracted enormous following-up investigations in the community of condensed matter physics and material sciences. But quite soon, the experimental results from different groups indicate consistently that no evidence of near-ambient superconductivity is found in the samples synthesized by the same method as the reported one, or by the other alternative methods. From our extended high-pressure heat capacity and magnetic susceptibility measurements on the samples prepared with the lutetium foil and H2/N2 gas mixture, we report the finding of a magnetic transition at the temperature about 56 K. Our results show that this magnetic phase is robust against pressure up to 4.3 GPa, which covers the critical pressure of boosting the claimed near room temperature superconductivity., 14 pages, 4 figures

Recently, text-guided 3D generative methods have made remarkable advancements in producing high-quality textures and geometry, capitalizing on the proliferation of large vision-language and image diffusion models. However, existing methods still struggle to create high-fidelity 3D head avatars in two aspects: (1) They rely mostly on a pre-trained text-to-image diffusion model whilst missing the necessary 3D awareness and head priors. This makes them prone to inconsistency and geometric distortions in the generated avatars. (2) They fall short in fine-grained editing. This is primarily due to the inherited limitations from the pre-trained 2D image diffusion models, which become more pronounced when it comes to 3D head avatars. In this work, we address these challenges by introducing a versatile coarse-to-fine pipeline dubbed HeadSculpt for crafting (i.e., generating and editing) 3D head avatars from textual prompts. Specifically, we first equip the diffusion model with 3D awareness by leveraging landmark-based control and a learned textual embedding representing the back view appearance of heads, enabling 3D-consistent head avatar generations. We further propose a novel identity-aware editing score distillation strategy to optimize a textured mesh with a high-resolution differentiable rendering technique. This enables identity preservation while following the editing instruction. We showcase HeadSculpt's superior fidelity and editing capabilities through comprehensive experiments and comparisons with existing methods., Webpage: https://brandonhan.uk/HeadSculpt/

Generative modelling over continuous-time geometric constructs, a.k.a chirographic data such as handwriting, sketches, drawings etc., have been accomplished through autoregressive distributions. Such strictly-ordered discrete factorization however falls short of capturing key properties of chirographic data -- it fails to build holistic understanding of the temporal concept due to one-way visibility (causality). Consequently, temporal data has been modelled as discrete token sequences of fixed sampling rate instead of capturing the true underlying concept. In this paper, we introduce a powerful model-class namely Denoising\ Diffusion\ Probabilistic\ Models or DDPMs for chirographic data that specifically addresses these flaws. Our model named "ChiroDiff", being non-autoregressive, learns to capture holistic concepts and therefore remains resilient to higher temporal sampling rate up to a good extent. Moreover, we show that many important downstream utilities (e.g. conditional sampling, creative mixing) can be flexibly implemented using ChiroDiff. We further show some unique use-cases like stochastic vectorization, de-noising/healing, abstraction are also possible with this model-class. We perform quantitative and qualitative evaluation of our framework on relevant datasets and found it to be better or on par with competing approaches.

A growing body of epidemiological evidence suggests that nonalcoholic fatty liver disease (NAFLD) is an independent risk factor for chronic kidney disease (CKD), but the regulatory mechanism linking NAFLD and CKD remains unclear. Our previous studies have shown that overexpression of PDE4D in mouse liver is sufficient for NAFLD, but little is known about its role in kidney injury. Here, liver-specific PDE4D conditional knockout (LKO) mice, adeno-associated virus 8 (AAV8)-mediated gene transfer of PDE4D and the PDE4 inhibitor roflumilast were used to assess the involvement of hepatic PDE4D in NAFLD-associated renal injury. We found that mice fed a high-fat diet (HFD) for 16 weeks developed hepatic steatosis and kidney injury, with an associated increase in hepatic PDE4D but no changes in renal PDE4D. Furthermore, liver-specific knockout of PDE4D or pharmacological inhibition of PDE4 with roflumilast ameliorated hepatic steatosis and kidney injury in HFD-fed diabetic mice. Correspondingly, overexpression of hepatic PDE4D resulted in significant renal damage. Mechanistically, highly expressed PDE4D in fatty liver promoted the production and secretion of TGF-β1 into blood, which triggered kidney injury by activating SMADs and subsequent collagen deposition. Our findings revealed PDE4D might act as a critical mediator between NAFLD and associated kidney injury and indicated PDE4 inhibitor roflumilast as a potential therapeutic strategy for NAFLD-associated CKD.

Generative modelling over continuous-time geometric constructs, a.k.a such as handwriting, sketches, drawings etc., have been accomplished through autoregressive distributions. Such strictly-ordered discrete factorization however falls short of capturing key properties of chirographic data -- it fails to build holistic understanding of the temporal concept due to one-way visibility (causality). Consequently, temporal data has been modelled as discrete token sequences of fixed sampling rate instead of capturing the true underlying concept. In this paper, we introduce a powerful model-class namely "Denoising Diffusion Probabilistic Models" or DDPMs for chirographic data that specifically addresses these flaws. Our model named "ChiroDiff", being non-autoregressive, learns to capture holistic concepts and therefore remains resilient to higher temporal sampling rate up to a good extent. Moreover, we show that many important downstream utilities (e.g. conditional sampling, creative mixing) can be flexibly implemented using ChiroDiff. We further show some unique use-cases like stochastic vectorization, de-noising/healing, abstraction are also possible with this model-class. We perform quantitative and qualitative evaluation of our framework on relevant datasets and found it to be better or on par with competing approaches., Accepted at ICLR '23

Background: Colorectal cancer (CRC) is the most common primary malignancy. Homoharringtonine (HHT) has potential therapeutic effects on solid tumors. However, the regulatory target and mechanism of HHT in CRC progression remain elusive. Methods: CCK-8, Edu staining, flow cytometry and Western blotting assay to detect the effect of HHT on proliferation and apoptosis of CRC cells. Transcriptome sequencing screened the drug delivery target of HHT. In vitro rescue experiment and in vivo tumorigenesis experiment detect the targeted interaction between HHT and NKD1. Quantitative proteomics combined with CO-IP and IF experiments to verify the targeted interaction between NKD1 and PCM1. Results: HHT suppressed CRC cells proliferation by inducing cell cycle arrest and apoptosis in vitro and vivo. HHT inhibited NKD1 expression in a concentration and time dependent manner. NKD1 was overexpressed in CRC and its depletion enhanced the therapeutic sensitivity of HHT on CRC, which indicating that NKD1 plays an important role in the development of CRC as the drug delivery target of HHT. Furthermore, proteomic analysis revealed that PCM1 participated the process of NKD1-regulated cell proliferation and cell cycle. NKD1 interacted with PCM1 and promotes PCM1 degradation through the ubiquitin-proteasome pathway. The overexpression of PCM1 effectively reversed the inhibition of siNKD1 on cell cycle. Conclusions: These results suggested that NKD1 targeting PCM1 to participate in regulating the therapeutic effects of HHT on CRC. Our findings provide evidence for clinical application of NKD1-targeted therapy in improving HHT sensitivity for CRC treatment

A material described as lutetium-hydrogen-nitrogen (Lu-H-N in short) was recently claimed to have near-ambient superconductivity[Gammon et al, Nature 615, 244, 2023]. If the results could be reproduced by other teams, it would be a major scientific breakthrough. Here, we report our results of transport and structure measurements on a material prepared using the same method as that reported by Gammon et al. Our X-ray diffraction measurements indicated that the obtained sample contained three substances: the FCC-1 phase (Fm-3m) with a lattice parameter a=5.03 {\AA}, the FCC-2 phase (Fm-3m) with a lattice parameter a= 4.755 {\AA} and Lu metal. These two FCC phases are identical to the those reported in the so-called near-ambient superconductor. However, we found that the samples had no evidence of superconductivity, through our resistance measurements in the temperature range of 300 - 4 K and pressure range of 0.9 - 3.4 GPa, and our magnetic susceptibility measurements in the pressure range of 0.8-3.3 GPa and temperature down to 100 K. We also used a laser heating technique to heat the sample to 1800{\deg}C and found no superconductivity in the produced dark blue samples below 6.5 GPa. In addition, the color of the both samples remain dark blue in the pressure range investigated., Comment: 13 pages and 4 figures

Background and Purpose: Peripheral nerve trauma-induced dysregulation of pain-associated genes in the primary sensory neurons of dorsal root ganglion (DRG) contributes to neuropathic pain genesis. RNA-binding proteins participate in gene transcription. We hypothesized that RALY, an RNA-binding protein, participated in nerve trauma-induced dysregulation of DRG pain-associated genes and nociceptive hypersensitivity. Methods and results: Immunohistochemistry staining showed that RALY was expressed exclusively in the nuclei of DRG neurons. Peripheral nerve trauma caused by chronic constriction injury (CCI) of unilateral sciatic nerve produced time-dependent increases in the levels of Raly mRNA and RALY protein in injured DRG. Blocking this increase through DRG microinjection of adeno-associated virus 5 (AAV5)-expressing Raly shRNA reduced the CCI-induced elevation in the amount of eukaryotic initiation factor 4 gamma 2 (eIF4G2) mRNA and eIF4G2 protein in injured DRG and mitigated the development and maintenance of CCI-induced nociceptive hypersensitivity, without altering basal (acute) response to noxious stimuli and locomotor activity. Mimicking DRG increased RALY through DRG microinjection of AAV5 expressing Raly mRNA upregulated the expression of eIF4G2 mRNA and eIF4G2 protein in the DRG and led to hypersensitive responses to noxious stimuli in the absence of nerve trauma. Mechanistically, CCI promoted the binding of RALY to the promoter of eIF4G2 gene and triggered its transcriptional activity. Conclusion and Implications: Our findings indicate that RALY participates in nerve trauma-induced nociceptive hypersensitivity likely through transcriptionally triggering eIF4G2 expression in the DRG. RALY may be a potential target in neuropathic pain management.

Nerve injury to peripheral somatosensory system causes refractory neuropathic pain. Maladaptive changes of gene expression in primary sensory neurons are considered molecular basis of this disorder. Long non-coding RNAs (lncRNAs) are key regulators of gene transcription; however, their significance in neuropathic pain remains largely elusive.Here, we reported a novel lncRNA, named sensory neuron-specific lncRNA (SS-lncRNA), for its expression exclusively in dorsal root ganglion (DRG) and trigeminal ganglion. SS-lncRNA was predominantly expressed in small DRG neurons and significantly downregulated due to a reduction of early B cell transcription factor 1 in injured DRG after nerve injury. Rescuing this downregulation reversed a decrease of the calcium-activated potassium channel subfamily N member 1 (KCNN1) in injured DRG and alleviated nerve injury-induced nociceptive hypersensitivity. Conversely, DRG downregulation of SS-lncRNA reduced the expression of KCNN1, decreased total potassium currents and afterhyperpolarization currents and increased excitability in DRG neurons and produced neuropathic pain symptoms.Mechanistically, downregulated SS-lncRNA resulted in the reductions of its binding to Kcnn1 promoter and heterogeneous nuclear ribonucleoprotein M (hnRNPM), consequent recruitment of less hnRNPM to the Kcnn1 promoter and silence of Kcnn1 gene transcription in injured DRG.These findings indicate that SS-lncRNA may relieve neuropathic pain through hnRNPM-mediated KCNN1 rescue in injured DRG and offer a novel therapeutic strategy specific for this disorder.

Aim. This study was aimed at investigating the mechanism of PATZ1 inducing apoptosis through PUMA in glioblastoma. Overexpressed PATZ1 was transfected to explore its role in inducing apoptosis in glioblastoma cells. Methods. The expression of protein was detected by western blotting assay. qRT-PCR assay was used to detect the expression of RNA. Confocal microscopy was used to analyze the correlation between PATZ1 and PUMA. TUNEL assay was used to detect the cell apoptosis. The ability of cell proliferation was detected by MTT assay and EDU assay. The effects of PATZ1 on cell apoptosis and tumor proliferation were observed in vivo by tumor xenograft mouse model. Results. The results showed that low PATZ1 expression correlates with poor prognosis in glioblastoma patients. Overexpression of PATZ1 inhibits glioma cell proliferation and induces apoptosis by activating intrinsic apoptotic pathways. PATZ1 colocalizes intracellularly with PUMA inducing apoptosis through PUMA in glioblastoma. Conclusion. PATZ1 plays a biological regulatory role in inducing apoptosis in glioblastoma, and this regulatory effect is related to PUMA, and the specific mechanism remains to be further explored.

ObjectivesWe investigated HPV genotypes in a large cohort of patients with definitive cervical histologic diagnosis.MethodsHPV testing was performed by real-time PCR assay, including 18 high-risk HPV (hrHPV) and 3 low-risk HPV (lrHPV). Totally 61,422 patients with documented HPV genotyping results within 6 months before cervical histologic diagnoses were included.ResultsHrHPV positive rate was 55.1% among all tested cases with the highest in adenosquamous carcinoma (94.1%), followed by squamous cell carcinoma (SCC) (93.7%), cervical intraepithelial neoplasia 2/3 (CIN2/3) (92.8%). HrHPV positive rates were significantly higher in high-grade squamous lesions than in those in glandular lesions. HPV16 was the most common genotype followed by HPV52 and HPV58 in CIN2/3. The most frequent hrHPV genotype in adenocarcinoma in situ (AIS) was HPV18, followed by HPV16, HPV45 and HPV52. In SCC cases, HPV16 was the most common type followed by HPV58, HPV52, HPV18 and HPV33. However, HPV18 showed significantly higher prevalence in adenocarcinoma and adenosquamous carcinoma than in that in SCC. Theoretically, the protective rates of 2/4-valent and 9-valent vaccine were 69.1% and 85.8% for cervical cancers.ConclusionsThe prevalence of HPV genotypes in Chinese population was different from that in Western population. Some hrHPV types were identified in cervical precancerous lesions and cancers, which are not included in current HPV vaccines. These data provide baseline knowledge for future HPV vaccine development.

Sketches are highly expressive, inherently capturing subjective and fine-grained visual cues. The exploration of such innate properties of human sketches has, however, been limited to that of image retrieval. In this paper, for the first time, we cultivate the expressiveness of sketches but for the fundamental vision task of object detection. The end result is a sketch-enabled object detection framework that detects based on what \textit{you} sketch -- \textit{that} ``zebra'' (e.g., one that is eating the grass) in a herd of zebras (instance-aware detection), and only the \textit{part} (e.g., ``head" of a ``zebra") that you desire (part-aware detection). We further dictate that our model works without (i) knowing which category to expect at testing (zero-shot) and (ii) not requiring additional bounding boxes (as per fully supervised) and class labels (as per weakly supervised). Instead of devising a model from the ground up, we show an intuitive synergy between foundation models (e.g., CLIP) and existing sketch models build for sketch-based image retrieval (SBIR), which can already elegantly solve the task -- CLIP to provide model generalisation, and SBIR to bridge the (sketch$\rightarrow$photo) gap. In particular, we first perform independent prompting on both sketch and photo branches of an SBIR model to build highly generalisable sketch and photo encoders on the back of the generalisation ability of CLIP. We then devise a training paradigm to adapt the learned encoders for object detection, such that the region embeddings of detected boxes are aligned with the sketch and photo embeddings from SBIR. Evaluating our framework on standard object detection datasets like PASCAL-VOC and MS-COCO outperforms both supervised (SOD) and weakly-supervised object detectors (WSOD) on zero-shot setups. Project Page: \url{https://pinakinathc.github.io/sketch-detect}, Accepted as Top 12 Best Papers. Will be presented in special single-track plenary sessions to all attendees in Computer Vision and Pattern Recognition (CVPR), 2023. Project Page: www.pinakinathc.me/sketch-detect

Human sketch has already proved its worth in various visual understanding tasks (e.g., retrieval, segmentation, image-captioning, etc). In this paper, we reveal a new trait of sketches - that they are also salient. This is intuitive as sketching is a natural attentive process at its core. More specifically, we aim to study how sketches can be used as a weak label to detect salient objects present in an image. To this end, we propose a novel method that emphasises on how "salient object" could be explained by hand-drawn sketches. To accomplish this, we introduce a photo-to-sketch generation model that aims to generate sequential sketch coordinates corresponding to a given visual photo through a 2D attention mechanism. Attention maps accumulated across the time steps give rise to salient regions in the process. Extensive quantitative and qualitative experiments prove our hypothesis and delineate how our sketch-based saliency detection model gives a competitive performance compared to the state-of-the-art., CVPR 2023. Project page available at https://ayankumarbhunia.github.io/Sketch2Saliency/

Background Left-behind adolescents (LBAs) are adolescents aged 11–18 years who are separated from their parents and left behind in local cities by one or both parents for a period of more than 6 months. LBAs in rural areas are likely to engage in aggressive behavior, which can affect interpersonal relationships, reduce academic performance, and even lead to anxiety and depression. To our knowledge, no studies have examined the mediating effect of resilience and self-esteem on the relationship between negative life events and aggression among Chinese rural LBAs. Therefore, this study aimed to explore the relationship between negative life events and aggression among Chinese rural LBAs and how self-esteem and resilience mediate the association. Methods Using a stratified random sampling method, 1344 LBAs in Hunan Province of China were investigated. Information was collected by a self-designed sociodemographic questionnaire, Adolescent Self-Rating Life Events Checklist, Resilience Scale Chinese Adolescent, Rosenberg Self-Esteem Scale and Aggression Scales to assess the psychology of LBAs. Data analysis was conducted using descriptive statistics, Pearson correlation, and regression analysis to estimate direct and indirect effects using bootstrap analysis. Results Negative life events were significantly related to self-esteem (r = − 0.338), resilience (r = − 0.359), and aggression (r = 0.441). Aggression was directly affected by self-esteem (β = − 0.44) and resilience (β = − 0.34). Negative life events were not only directly related to aggression (β = 0.34, 95% CI: 0.275 ~ 0.398) but also showed an indirect effect on aggression through self-esteem and resilience. The direct effect, total effect and indirect effect of negative life events on aggression through self-esteem and resilience were 0.3364, 0.4344 and 0.0980, respectively. The mediating effect of self-esteem and resilience accounted for 22.56% of the relationship between negative life events and aggression. Conclusions We found that self-esteem and resilience mediated most negative life events on aggression. It is imperative for educators and families to improve LBAs’ self-esteem and resilience to reduce the occurrence of aggression. Future intervention studies should be designed to strengthen self-esteem and resilience.

Charge order has recently been revealed in infinite-layer nickelates as a broken symmetry state in addition to superconductivity. Its root and link to superconductivity remain elusive. Here, we report a resonant x-ray scattering measurement of the charge order in infinite-layer PrNiO2. By measuring the scattering intensity versus azimuthal angles, we resolve the intra-unit-cell symmetry of the charge order and discover that it has predominant d-wave orbital symmetry, strikingly resembling lightly doped cuprates. Our results shed new light on the nature of charge order in infinite-layer nickelates and may assist in establishing a common thread to understanding the role of charge order in infinite-layer nickelates and cuprates., 17 pages, 4 figures

Deep hiding, concealing secret information using Deep Neural Networks (DNNs), can significantly increase the embedding rate and improve the efficiency of secret sharing. Existing works mainly force on designing DNNs with higher embedding rates or fancy functionalities. In this paper, we want to answer some fundamental questions: how to increase and what determines the embedding rate of deep hiding. To this end, we first propose a novel Local Deep Hiding (LDH) scheme that significantly increases the embedding rate by hiding large secret images into small local regions of cover images. Our scheme consists of three DNNs: hiding, locating, and revealing. We use the hiding network to convert a secret image in a small imperceptible compact secret code that is embedded into a random local region of a cover image. The locating network assists the revealing process by identifying the position of secret codes in the stego image, while the revealing network recovers all full-size secret images from these identified local regions. Our LDH achieves an extremely high embedding rate, i.e., $16\times24$ bpp and exhibits superior robustness to common image distortions. We also conduct comprehensive experiments to evaluate our scheme under various system settings. We further quantitatively analyze the trade-off between the embedding rate and image quality with different image restoration algorithms.

As an essential technology underpinning trusted computing, the trusted execution environment (TEE) allows one to launch computation tasks on both on- and off-premises data while assuring confidentiality and integrity. This article provides a systematic review and comparison of TEE-based secure computation protocols. We first propose a taxonomy that classifies secure computation protocols into three major categories, namely secure outsourced computation, secure distributed computation and secure multi-party computation. To enable a fair comparison of these protocols, we also present comprehensive assessment criteria with respect to four aspects: setting, methodology, security and performance. Based on these criteria, we review, discuss and compare the state-of-the-art TEE-based secure computation protocols for both general-purpose computation functions and special-purpose ones, such as privacy-preserving machine learning and encrypted database queries. To the best of our knowledge, this article is the first survey to review TEE-based secure computation protocols and the comprehensive comparison can serve as a guideline for selecting suitable protocols for deployment in practice. Finally, we also discuss several future research directions and challenges.

IntroductionThe diagnosis and treatment of ankylosing spondylitis (AS) is a difficult task, especially in less developed countries without access to experts. To address this issue, a comprehensive artificial intelligence (AI) tool was created to help diagnose and predict the course of AS.MethodsIn this retrospective study, a dataset of 5389 pelvic radiographs (PXRs) from patients treated at a single medical center between March 2014 and April 2022 was used to create an ensemble deep learning (DL) model for diagnosing AS. The model was then tested on an additional 583 images from three other medical centers, and its performance was evaluated using the area under the receiver operating characteristic curve analysis, accuracy, precision, recall, and F1 scores. Furthermore, clinical prediction models for identifying high-risk patients and triaging patients were developed and validated using clinical data from 356 patients.ResultsThe ensemble DL model demonstrated impressive performance in a multicenter external test set, with precision, recall, and area under the receiver operating characteristic curve values of 0.90, 0.89, and 0.96, respectively. This performance surpassed that of human experts, and the model also significantly improved the experts' diagnostic accuracy. Furthermore, the model's diagnosis results based on smartphone-captured images were comparable to those of human experts. Additionally, a clinical prediction model was established that accurately categorizes patients with AS into high-and low-risk groups with distinct clinical trajectories. This provides a strong foundation for individualized care.DiscussionIn this study, an exceptionally comprehensive AI tool was developed for the diagnosis and management of AS in complex clinical scenarios, especially in underdeveloped or rural areas that lack access to experts. This tool is highly beneficial in providing an efficient and effective system of diagnosis and management.

Colorectal cancer (CRC) is the most common primary malignant tumor with a significantly higher incidence in the worldwide. Homoharringtonine (HHT) often used to treatment of acute leukemia. Recent research revealed it could be used for solid cancer therapy. However, the regulatory target and mechanism of HHT in CRC progression remain elusive. This study proved that HHT suppressed cell proliferation and promoted cell cycle arrest and apoptosis. Transcriptome sequence indicated that NKD1 was the target of HHT in CRC. HHT could suppress NKD1 expression in a concentration and time dependent manner. NKD1 was overexpressed in CRC tissues and depletion of NKD1 enhanced the therapeutic effect of HHT on CRC in vitro and vivo. Furthermore, proteomic analysis revealed that PCM1 involved in the process of cell proliferation and cell cycle regulated by NKD1. NKD1 interacts with PCM1, and NKD1 promotes the ubiquitination degradation of PCM1. Moreover, overexpression of PCM1 can effectively reverse the promoting effect of NKD1 interference on cell cycle arrest and apoptosis. These results suggested that the NKD1/PCM1 axis participated in mediating the therapeutic sensitivity of HHT to CRC. Our findings provide evidence for clinical application of NKD1-targeted therapy in improving HHT sensitivity for CRC treatment.

A long-standing topic in artificial intelligence is the effective recognition of patterns from noisy images. In this regard, the recent data-driven paradigm considers 1) improving the representation robustness by adding noisy samples in training phase (i.e., data augmentation) or 2) pre-processing the noisy image by learning to solve the inverse problem (i.e., image denoising). However, such methods generally exhibit inefficient process and unstable result, limiting their practical applications. In this paper, we explore a non-learning paradigm that aims to derive robust representation directly from noisy images, without the denoising as pre-processing. Here, the noise-robust representation is designed as Fractional-order Moments in Radon space (FMR), with also beneficial properties of orthogonality and rotation invariance. Unlike earlier integer-order methods, our work is a more generic design taking such classical methods as special cases, and the introduced fractional-order parameter offers time-frequency analysis capability that is not available in classical methods. Formally, both implicit and explicit paths for constructing the FMR are discussed in detail. Extensive simulation experiments and an image security application are provided to demonstrate the uniqueness and usefulness of our FMR, especially for noise robustness, rotation invariance, and time-frequency discriminability.

Here we report the first investigation on correlation between the transport properties and the corresponding stability of the lattice structure for CaK(Fe1-xNix)4As4 (x=0.04 and 0), a new type of putative topological superconductors, with and without a spin-vortex crystal (SVC) state in a wide pressure range involving superconducting to non-superconducting transition and the half- to full-collapse of tetragonal (h-cT and f-cT) phases, by the complementary measurements of high-pressure resistance, Hall coefficient and synchrotron X-ray diffraction. We identify the three critical pressures, P1 that is the turn-on critical pressure of the h-cT phase transition and it coincides with the critical pressure for the sign change of Hall coefficient from positive to negative, a manifestation of the Fermi surface reconstruction, P2 that is the turn-off pressures of the h-cT phase transition, and P3 that is the critical pressure of the f-cT phase transition. By comparing the high-pressure results measured from the two kinds of samples, we find a distinct left-shift of the P1 for the doped sample, at the pressure of which its SVC state is fully suppressed, however the P2 and the P3 remain the same as that of the undoped one. Our results not only provide a consistent understanding on the results reported before, but also demonstrate the importance of the Fe-As bonding in stabilizing the superconductivity of the iron pnictide superconductors through the pressure window.

BackgroundTuberculosis (TB) is a chronic infectious disease. Bone and joint TB is a common type of extrapulmonary TB and often occurs secondary to TB infection. In this study, we aimed to find the difference in the blood examination results of patients with bone and joint TB and patients with TB by using machine learning (ML) and establish a diagnostic model to help clinicians better diagnose the disease and allow patients to receive timely treatment.MethodsA total of 1,667 patients were finally enrolled in the study. Patients were randomly assigned to the training and validation cohorts. The training cohort included 1,268 patients: 158 patients with bone and joint TB and 1,110 patients with TB. The validation cohort included 399 patients: 48 patients with bone and joint TB and 351 patients with TB. We used three ML methods, namely logistic regression, LASSO regression, and random forest, to screen the differential variables, obtained the most representative variables by intersection to construct the prediction model, and verified the performance of the proposed prediction model in the validation group.ResultsThe results revealed a great difference in the blood examination results of patients with bone and joint TB and those with TB. Infectious markers such as hs-CRP, ESR, WBC, and NEUT were increased in patients with bone and joint TB. Patients with bone and joint TB were found to have higher liver function burden and poorer nutritional status. The factors screened using ML were PDW, LYM, AST/ALT, BUN, and Na, and the nomogram diagnostic model was constructed using these five factors. In the training cohort, the area under the curve (AUC) value of the model was 0.71182, and the C value was 0.712. In the validation cohort, the AUC value of the model was 0.6435779, and the C value was 0.644.ConclusionWe used ML methods to screen out the blood-specific factors—PDW, LYM, AST/ALT, BUN, and Na+—of bone and joint TB and constructed a diagnostic model to help clinicians better diagnose the disease in the future.

In this paper, we leverage CLIP for zero-shot sketch based image retrieval (ZS-SBIR). We are largely inspired by recent advances on foundation models and the unparalleled generalisation ability they seem to offer, but for the first time tailor it to benefit the sketch community. We put forward novel designs on how best to achieve this synergy, for both the category setting and the fine-grained setting ("all"). At the very core of our solution is a prompt learning setup. First we show just via factoring in sketch-specific prompts, we already have a category-level ZS-SBIR system that overshoots all prior arts, by a large margin (24.8%) - a great testimony on studying the CLIP and ZS-SBIR synergy. Moving onto the fine-grained setup is however trickier, and requires a deeper dive into this synergy. For that, we come up with two specific designs to tackle the fine-grained matching nature of the problem: (i) an additional regularisation loss to ensure the relative separation between sketches and photos is uniform across categories, which is not the case for the gold standard standalone triplet loss, and (ii) a clever patch shuffling technique to help establishing instance-level structural correspondences between sketch-photo pairs. With these designs, we again observe significant performance gains in the region of 26.9% over previous state-of-the-art. The take-home message, if any, is the proposed CLIP and prompt learning paradigm carries great promise in tackling other sketch-related tasks (not limited to ZS-SBIR) where data scarcity remains a great challenge. Project page: https://aneeshan95.github.io/Sketch_LVM/, Comment: Accepted in CVPR 2023. Project page available at https://aneeshan95.github.io/Sketch_LVM/

In the fashion domain, there exists a variety of vision-and-language (V+L) tasks, including cross-modal retrieval, text-guided image retrieval, multi-modal classification, and image captioning. They differ drastically in each individual input/output format and dataset size. It has been common to design a task-specific model and fine-tune it independently from a pre-trained V+L model (e.g., CLIP). This results in parameter inefficiency and inability to exploit inter-task relatedness. To address such issues, we propose a novel FAshion-focused Multi-task Efficient learning method for Vision-and-Language tasks (FAME-ViL) in this work. Compared with existing approaches, FAME-ViL applies a single model for multiple heterogeneous fashion tasks, therefore being much more parameter-efficient. It is enabled by two novel components: (1) a task-versatile architecture with cross-attention adapters and task-specific adapters integrated into a unified V+L model, and (2) a stable and effective multi-task training strategy that supports learning from heterogeneous data and prevents negative transfer. Extensive experiments on four fashion tasks show that our FAME-ViL can save 61.5% of parameters over alternatives, while significantly outperforming the conventional independently trained single-task models. Code is available at https://github.com/BrandonHanx/FAME-ViL., Comment: CVPR 2023

Searching for a universal trend by the same tuning method in different high-temperature superconductors with a similar crystal structure is a common strategy to find clues for a better understanding of the superconducting mechanism in a unified way. It is known that the hole-doped bismuth-oxide Ba1-xKxBiO3 possesses a similar perovskite structure to that of the hole-doped copper-oxide (cuprate) superconductors but also holds a comparatively high superconducting transition temperature. In this study, we report the first observation of the pressure-induced quantum phase transition (QPT) from superconducting to insulating states in a series of Ba1-xKxBiO3 single-crystal samples. A similar QPT has also been observed recently in the compressed cuprate superconductors1. Significantly, we found that the QPT observed in Ba1-xKxBiO3 is intriguingly associated with the valence change of the Bi ions in the material. These results lead us to propose that the pressure-induced valence change from Bi3+ to Bi5+ destroys the hole-doping effect on stabilizing the conductivity and corresponding superconductivity. By comparing the high-pressure behaviors observed in these two kinds of oxides, we identified another prominent feature shared by them - the more the hole-doping concentration, the higher the critical pressure required for driving the QPT., Comment: 20 pages and 4 figures

We investigate the electronic and phonon properties of hydrogen sulfide (SH$_3$) under ultrahigh pressure to elucidate the origin of its high-T$_c$ superconductivity. Contrary to the prevailing belief that the metalized S-H $\sigma$ bond is responsible, our analysis, based on the anisotropic Migdal-Eliashberg equation and the crystal orbital Hamilton population (COHP) calculation, reveals that the H-H $\sigma$-antibonding states play a dominant role in the large electron-phonon coupling that leads to the superconducting pairing in SH$_3$. Furthermore, by partially restricting the vibration of S atoms, we demonstrate that the S-H bonds provide subsidiary contributions to the pairing interaction. These findings shed light on the importance of the previously overlooked H-H $\sigma^*$ bonds in driving high-T$_c$ superconductivity in SH$_3$ and offer insights into the relationship between metallic H-H covalent antibonding and high-T$_c$ superconductivity in other hydrogen-rich materials under high pressure., Comment: 11 pages, 9 figures, 59 references

Target domain pseudo-labelling has shown effectiveness in unsupervised domain adaptation (UDA). However, pseudo-labels of unlabeled target domain data are inevitably noisy due to the distribution shift between source and target domains. This paper proposes a Generative model-based Noise-Robust Training method (GeNRT), which eliminates domain shift while mitigating label noise. GeNRT incorporates a Distribution-based Class-wise Feature Augmentation (D-CFA) and a Generative-Discriminative classifier Consistency (GDC), both based on the class-wise target distributions modelled by generative models. D-CFA minimizes the domain gap by augmenting the source data with distribution-sampled target features, and trains a noise-robust discriminative classifier by using target domain knowledge from the generative models. GDC regards all the class-wise generative models as generative classifiers and enforces a consistency regularization between the generative and discriminative classifiers. It exploits an ensemble of target knowledge from all the generative models to train a noise-robust discriminative classifier and eventually gets theoretically linked to the Ben-David domain adaptation theorem for reducing the domain gap. Extensive experiments on Office-Home, PACS, and Digit-Five show that our GeNRT achieves comparable performance to state-of-the-art methods under single-source and multi-source UDA settings.

We propose a new formulation of temporal action detection (TAD) with denoising diffusion, DiffTAD in short. Taking as input random temporal proposals, it can yield action proposals accurately given an untrimmed long video. This presents a generative modeling perspective, against previous discriminative learning manners. This capability is achieved by first diffusing the ground-truth proposals to random ones (i.e., the forward/noising process) and then learning to reverse the noising process (i.e., the backward/denoising process). Concretely, we establish the denoising process in the Transformer decoder (e.g., DETR) by introducing a temporal location query design with faster convergence in training. We further propose a cross-step selective conditioning algorithm for inference acceleration. Extensive evaluations on ActivityNet and THUMOS show that our DiffTAD achieves top performance compared to previous art alternatives. The code will be made available at https://github.com/sauradip/DiffusionTAD., Comment: ICCV 2023; Code available at https://github.com/sauradip/DiffusionTAD

This paper advances the fine-grained sketch-based image retrieval (FG-SBIR) literature by putting forward a strong baseline that overshoots prior state-of-the-arts by ~11%. This is not via complicated design though, but by addressing two critical issues facing the community (i) the gold standard triplet loss does not enforce holistic latent space geometry, and (ii) there are never enough sketches to train a high accuracy model. For the former, we propose a simple modification to the standard triplet loss, that explicitly enforces separation amongst photos/sketch instances. For the latter, we put forward a novel knowledge distillation module can leverage photo data for model training. Both modules are then plugged into a novel plug-n-playable training paradigm that allows for more stable training. More specifically, for (i) we employ an intra-modal triplet loss amongst sketches to bring sketches of the same instance closer from others, and one more amongst photos to push away different photo instances while bringing closer a structurally augmented version of the same photo (offering a gain of ~4-6%). To tackle (ii), we first pre-train a teacher on the large set of unlabelled photos over the aforementioned intra-modal photo triplet loss. Then we distill the contextual similarity present amongst the instances in the teacher's embedding space to that in the student's embedding space, by matching the distribution over inter-feature distances of respective samples in both embedding spaces (delivering a further gain of ~4-5%). Apart from outperforming prior arts significantly, our model also yields satisfactory results on generalising to new classes. Project page: https://aneeshan95.github.io/Sketch_PVT/, Comment: Accepted in CVPR 2023. Project page available at https://aneeshan95.github.io/Sketch_PVT/

Virtual try-on, i.e making people virtually try new garments, is an active research area in computer vision with great commercial applications. Current virtual try-on methods usually work in a two-stage pipeline. First, the garment image is warped on the person's pose using a flow estimation network. Then in the second stage, the warped garment is fused with the person image to render a new try-on image. Unfortunately, such methods are heavily dependent on the quality of the garment warping which often fails when dealing with hard poses (e.g., a person lifting or crossing arms). In this work, we propose a robust warping method for virtual try-on based on a learned garment DensePose which has a direct correspondence with the person's DensePose. Due to the lack of annotated data, we show how to leverage an off-the-shelf person DensePose model and a pretrained flow model to learn the garment DensePose in a weakly supervised manner. The garment DensePose allows a robust warping to any person's pose without any additional computation. Our method achieves the state-of-the-art equivalent on virtual try-on benchmarks and shows warping robustness on in-the-wild person images with hard poses, making it more suited for real-world virtual try-on applications., Comment: 6 pages