FCPN: Pruning redundant part-whole relations for more streamlined pattern parsing.

Cropping-and-segmenting pattern parsers often combine diverse inner correlations into a single metric/scheme, resulting in over-generalizations and redundant representations. It is proposed to streamline pattern parsing by using presenting a redundant association elimination network (RAEN) with capsule attention twisters (CATs) and capsule-attention routing agreement (CARA). CATs trim delicate relationships between parts and wholes that are weak and interchangeable. Senior entities can only be updated by primary entities that meet the requirements of inter-part diversity and intra-object cohesiveness. In order to enhance results, CARA is designed to protect against the unnecessary voting signals of traditional routing protocols. Experiments involving facial and human segmentation show that RAEN is better than current remarkable methods, particularly for defining detailed semantic boundaries.

FCPN: Pruning redundant part-whole relations for more streamlined pattern parsing.

Most cropping-and-segmenting pattern parsers typically establish a single metric/scheme to reason diverse inner correlations, resulting in over-general and redundant representations. To make pattern parsing more streamlined and efficient, a fragile correlation pruner network (FCPN) with correlation-steered attention shifters (CSASs) and graph attention expectation-maximum routing agreement (GAEMRA) is proposed. CSASs prune fragile (weak and substitutable) part-to-whole correlations. They stipulate that only those primary entities (representing components) fulfilling the criteria of inter-part diversity and intra-object cohesiveness can update senior entities (representing the whole/intermediate composites). To further boost effects, GAEMRA is defined to shield the redundant voting signals of conventional routing agreement. With CSASs and GAEMRA, FCPN gradually parses objective semantic patterns by clustering highly associated secondary entities in a bottom-up "part backtracking" manner. Quantitative and ablation experiments surrounding face and human parsing demonstrate the superiority of FCPN over the state-of-the-arts, especially for the definition of fine-grained semantic boundaries.

Organoid modeling of lung-resident immune responses to SARS-CoV-2 infection.

Tissue-resident immunity underlies essential host defenses against pathogens, but analysis in humans has lacked in vitro model systems where epithelial infection and accompanying resident immune cell responses can be observed en bloc. Indeed, human primary epithelial organoid cultures typically omit immune cells, and human tissue resident-memory lymphocytes are conventionally assayed without an epithelial infection component, for instance from peripheral blood, or after extraction from organs. Further, the study of resident immunity in animals can be complicated by interchange between tissue and peripheral immune compartments. To study human tissue-resident infectious immune responses in isolation from secondary lymphoid organs, we generated adult human lung three-dimensional air-liquid interface (ALI) lung organoids from intact tissue fragments that co-preserve epithelial and stromal architecture alongside endogenous lung-resident immune subsets. These included T, B, NK and myeloid cells, with CD69+CD103+ tissue-resident and CCR7- and/or CD45RA- TRM and conservation of T cell receptor repertoires, all corresponding to matched fresh tissue. SARS-CoV-2 vigorously infected organoid lung epithelium, alongside secondary induction of innate cytokine production that was inhibited by antiviral agents. Notably, SARS-CoV-2-infected organoids manifested adaptive virus-specific T cell activation that was specific for seropositive and/or previously infected donor individuals. This holistic non-reconstitutive organoid system demonstrates the sufficiency of lung to autonomously mount adaptive T cell memory responses without a peripheral lymphoid component, and represents an enabling method for the study of human tissue-resident immunity.

IRX5 promotes DNA damage repair and activation of hair follicle stem cells.

The molecular mechanisms allowing hair follicles to periodically activate their stem cells (HFSCs) are incompletely characterized. Here, we identify the transcription factor IRX5 as a promoter of HFSC activation. Irx5-/- mice have delayed anagen onset, with increased DNA damage and diminished HFSC proliferation. Open chromatin regions form near cell cycle progression and DNA damage repair genes in Irx5-/- HFSCs. DNA damage repair factor BRCA1 is an IRX5 downstream target. Inhibition of FGF kinase signaling partially rescues the anagen delay in Irx5-/- mice, suggesting that the Irx5-/- HFSC quiescent phenotype is partly due to failure to suppress Fgf18 expression. Interfollicular epidermal stem cells also show decreased proliferation and increased DNA damage in Irx5-/-mice. Consistent with a role for IRX5 as a promoter of DNA damage repair, we find that IRX genes are upregulated in many cancer types and that there is a correlation between IRX5 and BRCA1 expression in breast cancer.

Feature Correlation-Steered Capsule Network for object detection.

Despite Convolutional Neural Networks (CNNs) based approaches have been successful in objects detection, they predominantly focus on positioning discriminative regions while overlooking the internal holistic part-whole associations within objects. This would ultimately lead to the neglect of feature relationships between object and its parts as well as among those parts, both of which are significantly helpful for detecting discriminative parts. In this paper, we propose to "look insider the objects" by digging into part-whole feature correlations and take the attempts to leverage those correlations endowed by the Capsule Network (CapsNet) for robust object detection. Actually, highly correlated capsules across adjacent layers share high familiarity, which will be more likely to be routed together. In light of this, we take such correlations between different capsules of the preceding training samples as an awareness to constrain the subsequent candidate voting scope during the routing procedure, and a Feature Correlation-Steered CapsNet (FCS-CapsNet) with Locally-Constrained Expectation-Maximum (EM) Routing Agreement (LCEMRA) is proposed. Different from conventional EM routing, LCEMRA stipulates that only those relevant low-level capsules (parts) meeting the requirement of quantified intra-object cohesiveness can be clustered to make up high-level capsules (objects). In doing so, part-object associations can be dug by transformation weighting matrixes between capsules layers during such "part backtracking" procedure. LCEMRA enables low-level capsules to selectively gather projections from a non-spatially-fixed set of high-level capsules. Experiments on VOC2007, VOC2012, HKU-IS, DUTS, and COCO show that FCS-CapsNet can achieve promising object detection effects across multiple evaluation metrics, which are on-par with state-of-the-arts.

Murine interfollicular epidermal differentiation is gradualistic with GRHL3 controlling progression from stem to transition cell states.

The interfollicular epidermis (IFE) forms a water-tight barrier that is often disrupted in inflammatory skin diseases. During homeostasis, the IFE is replenished by stem cells in the basal layer that differentiate as they migrate toward the skin surface. Conventionally, IFE differentiation is thought to be stepwise as reflected in sharp boundaries between its basal, spinous, granular and cornified layers. The transcription factor GRHL3 regulates IFE differentiation by transcriptionally activating terminal differentiation genes. Here we use single cell RNA-seq to show that murine IFE differentiation is best described as a single step gradualistic process with a large number of transition cells between the basal and spinous layer. RNA-velocity analysis identifies a commitment point that separates the plastic basal and transition cell state from unidirectionally differentiating cells. We also show that in addition to promoting IFE terminal differentiation, GRHL3 is essential for suppressing epidermal stem cell expansion and the emergence of an abnormal stem cell state by suppressing Wnt signaling in stem cells.