Instance segmentation in electron microscopy (EM) volumes poses a significant challenge due to the complex morphology of instances and insufficient annotations. Self-supervised learning has recently emerged as a promising solution, enabling the acquisition of prior knowledge of cellular tissue structures that are essential for EM instance segmentation. However, existing pretraining methods often lack the ability to capture complex visual patterns and relationships between voxels, which results in the acquired prior knowledge being insufficient for downstream EM analysis tasks. In this paper, we propose a novel pretraining framework that leverages multiscale visual representations to capture both voxel-level and feature-level consistency in EM volumes. Specifically, our framework enforces voxel-level consistency between the outputs of a Siamese network by a reconstruction function, and incorporates a cross-attention mechanism for soft feature matching to achieve fine-grained feature-level consistency. Moreover, we propose a contrastive learning scheme on the feature pyramid to extract discriminative features across multiple scales. We extensively pretrain our method on four large-scale EM datasets, achieving promising performance improvements in representative tasks of neuron and mitochondria instance segmentation.
Self-supervised sound source localization is usually challenged by the modality inconsistency. In recent studies, contrastive learning based strategies have shown promising to establish such a consistent correspondence between audio and sound sources in visual scenarios. Unfortunately, the insufficient attention to the heterogeneity influence in the different modality features still limits this scheme to be further improved, which also becomes the motivation of our work. In this study, an Induction Network is proposed to bridge the modality gap more effectively. By decoupling the gradients of visual and audio modalities, the discriminative visual representations of sound sources can be learned with the designed Induction Vector in a bootstrap manner, which also enables the audio modality to be aligned with the visual modality consistently. In addition to a visual weighted contrastive loss, an adaptive threshold selection strategy is introduced to enhance the robustness of the Induction Network. Substantial experiments conducted on SoundNet-Flickr and VGG-Sound Source datasets have demonstrated a superior performance compared to other state-of-the-art works in different challenging scenarios. The code is available at https://github.com/Tahy1/AVIN
Recent stereo matching networks achieves dramatic performance by introducing epipolar line constraint to limit the matching range of dual-view. However, in complicated real-world scenarios, the feature information based on intra-epipolar line alone is too weak to facilitate stereo matching. In this paper, we present a simple but highly effective network called Alternated Attention U-shaped Transformer (AAUformer) to balance the impact of epipolar line in dual and single view respectively for excellent generalization performance. Compared to other models, our model has several main designs: 1) to better liberate the local semantic features of the single-view at pixel level, we introduce window self-attention to break the limits of intra-row self-attention and completely replace the convolutional network for denser features before cross-matching; 2) the multi-scale alternated attention backbone network was designed to extract invariant features in order to achieves the coarse-to-fine matching process for hard-to-discriminate regions. We performed a series of both comparative studies and ablation studies on several mainstream stereo matching datasets. The results demonstrate that our model achieves state-of-the-art on the Scene Flow dataset, and the fine-tuning performance is competitive on the KITTI 2015 dataset. In addition, for cross generalization experiments on synthetic and real-world datasets, our model outperforms several state-of-the-art works.
Machine learning has made remarkable advancements, but confidently utilising learning-enabled components in safety-critical domains still poses challenges. Among the challenges, it is known that a rigorous, yet practical, way of achieving safety guarantees is one of the most prominent. In this paper, we first discuss the engineering and research challenges associated with the design and verification of such systems. Then, based on the observation that existing works cannot actually achieve provable guarantees, we promote a two-step verification method for the ultimate achievement of provable statistical guarantees.
Graph collaborative filtering (GCF) has gained considerable attention in recommendation systems by leveraging graph learning techniques to enhance collaborative filtering (CF) models. One classical approach in GCF is to learn user and item embeddings by modeling complex graph relations and utilizing these embeddings for CF models. However, the quality of the embeddings significantly impacts the recommendation performance of GCF models. In this paper, we argue that existing graph learning methods are insufficient in generating satisfactory embeddings for CF models. This is because they aggregate neighboring node messages directly, which can result in incorrect estimations of user-item correlations. To overcome this limitation, we propose a novel approach that incorporates causal modeling to explicitly encode the causal effects of neighboring nodes on the target node. This approach enables us to identify spurious correlations and uncover the root causes of user preferences. We introduce Causal Neural Graph Collaborative Filtering (CNGCF), the first causality-aware graph learning framework for CF. CNGCF integrates causal modeling into the graph representation learning process, explicitly coupling causal effects between node pairs into the core message-passing process of graph learning. As a result, CNGCF yields causality-aware embeddings that promote robust recommendations. Our extensive experiments demonstrate that CNGCF provides precise recommendations that align with user preferences. Therefore, our proposed framework can address the limitations of existing GCF models and offer a more effective solution for recommendation systems.
DeepFake based digital facial forgery is threatening public media security, especially when lip manipulation has been used in talking face generation, and the difficulty of fake video detection is further improved. By only changing lip shape to match the given speech, the facial features of identity are hard to be discriminated in such fake talking face videos. Together with the lack of attention on audio stream as the prior knowledge, the detection failure of fake talking face videos also becomes inevitable. It's found that the optical flow of the fake talking face video is disordered especially in the lip region while the optical flow of the real video changes regularly, which means the motion feature from optical flow is useful to capture manipulation cues. In this study, a fake talking face detection network (FTFDNet) is proposed by incorporating visual, audio and motion features using an efficient cross-modal fusion (CMF) module. Furthermore, a novel audio-visual attention mechanism (AVAM) is proposed to discover more informative features, which can be seamlessly integrated into any audio-visual CNN architecture by modularization. With the additional AVAM, the proposed FTFDNet is able to achieve a better detection performance than other state-of-the-art DeepFake video detection methods not only on the established fake talking face detection dataset (FTFDD) but also on the DeepFake video detection datasets (DFDC and DF-TIMIT).
Graph neural networks (GNNs) have pioneered advancements in graph representation learning, exhibiting superior feature learning and performance over multilayer perceptrons (MLPs) when handling graph inputs. However, understanding the feature learning aspect of GNNs is still in its initial stage. This study aims to bridge this gap by investigating the role of graph convolution within the context of feature learning theory in neural networks using gradient descent training. We provide a distinct characterization of signal learning and noise memorization in two-layer graph convolutional networks (GCNs), contrasting them with two-layer convolutional neural networks (CNNs). Our findings reveal that graph convolution significantly augments the benign overfitting regime over the counterpart CNNs, where signal learning surpasses noise memorization, by approximately factor $\sqrt{D}^{q-2}$, with $D$ denoting a node's expected degree and $q$ being the power of the ReLU activation function where $q > 2$. These findings highlight a substantial discrepancy between GNNs and MLPs in terms of feature learning and generalization capacity after gradient descent training, a conclusion further substantiated by our empirical simulations.
Vision-Language Pretraining (VLP) has demonstrated remarkable capabilities in learning visual representations from textual descriptions of images without annotations. Yet, effective VLP demands large-scale image-text pairs, a resource that suffers scarcity in the medical domain. Moreover, conventional VLP is limited to 2D images while medical images encompass diverse modalities, often in 3D, making the learning process more challenging. To address these challenges, we present Generative Text-Guided 3D Vision-Language Pretraining for Unified Medical Image Segmentation (GTGM), a framework that extends of VLP to 3D medical images without relying on paired textual descriptions. Specifically, GTGM utilizes large language models (LLM) to generate medical-style text from 3D medical images. This synthetic text is then used to supervise 3D visual representation learning. Furthermore, a negative-free contrastive learning objective strategy is introduced to cultivate consistent visual representations between augmented 3D medical image patches, which effectively mitigates the biases associated with strict positive-negative sample pairings. We evaluate GTGM on three imaging modalities - Computed Tomography (CT), Magnetic Resonance Imaging (MRI), and electron microscopy (EM) over 13 datasets. GTGM's superior performance across various medical image segmentation tasks underscores its effectiveness and versatility, by enabling VLP extension into 3D medical imagery while bypassing the need for paired text.
In this paper, we study the (geospatial) ontologies we are interested in together as an ontology (a geospatial ontology) system, consisting of a set of the (geospatial) ontologies and a set of ontology operations. A homomorphism between two ontology systems is a function between two sets of ontologies, which preserves these ontology operations. We view clustering a set of the ontologies we are interested in as partitioning the set or defining an equivalence relation on the set or forming a quotient set of the set or obtaining the surjective image of the set. Each ontology system homomorphism can be factored as a surjective clustering to a quotient space, followed by an embedding. Ontology (merging) systems, natural partial orders on the systems, and ontology merging closures in the systems are then transformed under ontology system homomorphisms, given by quotients and embeddings.
Large Language Models (LLMs) have exploded a new heatwave of AI, for their ability to engage end-users in human-level conversations with detailed and articulate answers across many knowledge domains. In response to their fast adoption in many industrial applications, this survey concerns their safety and trustworthiness. First, we review known vulnerabilities of the LLMs, categorising them into inherent issues, intended attacks, and unintended bugs. Then, we consider if and how the Verification and Validation (V&V) techniques, which have been widely developed for traditional software and deep learning models such as convolutional neural networks, can be integrated and further extended throughout the lifecycle of the LLMs to provide rigorous analysis to the safety and trustworthiness of LLMs and their applications. Specifically, we consider four complementary techniques: falsification and evaluation, verification, runtime monitoring, and ethical use. Considering the fast development of LLMs, this survey does not intend to be complete (although it includes 300 references), especially when it comes to the applications of LLMs in various domains, but rather a collection of organised literature reviews and discussions to support the quick understanding of the safety and trustworthiness issues from the perspective of V&V.