Clustering is a fundamental unsupervised representation learning task with wide application in computer vision and pattern recognition. Deep clustering utilizes deep neural networks to learn latent representation, which is suitable for clustering. However, previous deep clustering methods, especially image clustering, focus on the features of the data itself and ignore the relationship between the data, which is crucial for clustering. In this paper, we propose a novel Deep Structure and Attention aware Subspace Clustering (DSASC), which simultaneously considers data content and structure information. We use a vision transformer to extract features, and the extracted features are divided into two parts, structure features, and content features. The two features are used to learn a more efficient subspace structure for spectral clustering. Extensive experimental results demonstrate that our method significantly outperforms state-of-the-art methods. Our code will be available at https://github.com/cs-whh/DSASC
Low-light image enhancement aims to improve the perception of images collected in dim environments and provide high-quality data support for image recognition tasks. When dealing with photos captured under non-uniform illumination, existing methods cannot adaptively extract the differentiated luminance information, which will easily cause over-exposure and under-exposure. From the perspective of unsupervised learning, we propose a multi-scale attention Transformer named MSATr, which sufficiently extracts local and global features for light balance to improve the visual quality. Specifically, we present a multi-scale window division scheme, which uses exponential sequences to adjust the window size of each layer. Within different-sized windows, the self-attention computation can be refined, ensuring the pixel-level feature processing capability of the model. For feature interaction across windows, a global transformer branch is constructed to provide comprehensive brightness perception and alleviate exposure problems. Furthermore, we propose a loop training strategy, using the diverse images generated by weighted mixing and a luminance consistency loss to improve the model's generalization ability effectively. Extensive experiments on several benchmark datasets quantitatively and qualitatively prove that our MSATr is superior to state-of-the-art low-light image enhancement methods, and the enhanced images have more natural brightness and outstanding details. The code is released at https://github.com/fang001021/MSATr.
Environment shifts and conflicts present significant challenges for learning-based sound event localization and detection (SELD) methods. SELD systems, when trained in particular acoustic settings, often show restricted generalization capabilities for diverse acoustic environments. Furthermore, it is notably costly to obtain annotated samples for spatial sound events. Deploying a SELD system in a new environment requires extensive time for re-training and fine-tuning. To overcome these challenges, we propose environment-adaptive Meta-SELD, designed for efficient adaptation to new environments using minimal data. Our method specifically utilizes computationally synthesized spatial data and employs Model-Agnostic Meta-Learning (MAML) on a pre-trained, environment-independent model. The method then utilizes fast adaptation to unseen real-world environments using limited samples from the respective environments. Inspired by the Learning-to-Forget approach, we introduce the concept of selective memory as a strategy for resolving conflicts across environments. This approach involves selectively memorizing target-environment-relevant information and adapting to the new environments through the selective attenuation of model parameters. In addition, we introduce environment representations to characterize different acoustic settings, enhancing the adaptability of our attenuation approach to various environments. We evaluate our proposed method on the development set of the Sony-TAU Realistic Spatial Soundscapes 2023 (STARSS23) dataset and computationally synthesized scenes. Experimental results demonstrate the superior performance of the proposed method compared to conventional supervised learning methods, particularly in localization.
Continual learning (CL) has shown promising results and comparable performance to learning at once in a fully supervised manner. However, CL strategies typically require a large number of labeled samples, making their real-life deployment challenging. In this work, we focus on semi-supervised continual learning (SSCL), where the model progressively learns from partially labeled data with unknown categories. We provide a comprehensive analysis of SSCL and demonstrate that unreliable distributions of unlabeled data lead to unstable training and refinement of the progressing stages. This problem severely impacts the performance of SSCL. To address the limitations, we propose a novel approach called Dynamic Sub-Graph Distillation (DSGD) for semi-supervised continual learning, which leverages both semantic and structural information to achieve more stable knowledge distillation on unlabeled data and exhibit robustness against distribution bias. Firstly, we formalize a general model of structural distillation and design a dynamic graph construction for the continual learning progress. Next, we define a structure distillation vector and design a dynamic sub-graph distillation algorithm, which enables end-to-end training and adaptability to scale up tasks. The entire proposed method is adaptable to various CL methods and supervision settings. Finally, experiments conducted on three datasets CIFAR10, CIFAR100, and ImageNet-100, with varying supervision ratios, demonstrate the effectiveness of our proposed approach in mitigating the catastrophic forgetting problem in semi-supervised continual learning scenarios.
Recently, researchers have attempted to investigate the capability of LLMs in handling videos and proposed several video LLM models. However, the ability of LLMs to handle video grounding (VG), which is an important time-related video task requiring the model to precisely locate the start and end timestamps of temporal moments in videos that match the given textual queries, still remains unclear and unexplored in literature. To fill the gap, in this paper, we propose the LLM4VG benchmark, which systematically evaluates the performance of different LLMs on video grounding tasks. Based on our proposed LLM4VG, we design extensive experiments to examine two groups of video LLM models on video grounding: (i) the video LLMs trained on the text-video pairs (denoted as VidLLM), and (ii) the LLMs combined with pretrained visual description models such as the video/image captioning model. We propose prompt methods to integrate the instruction of VG and description from different kinds of generators, including caption-based generators for direct visual description and VQA-based generators for information enhancement. We also provide comprehensive comparisons of various VidLLMs and explore the influence of different choices of visual models, LLMs, prompt designs, etc, as well. Our experimental evaluations lead to two conclusions: (i) the existing VidLLMs are still far away from achieving satisfactory video grounding performance, and more time-related video tasks should be included to further fine-tune these models, and (ii) the combination of LLMs and visual models shows preliminary abilities for video grounding with considerable potential for improvement by resorting to more reliable models and further guidance of prompt instructions.
Existing learning-based hyperspectral reconstruction methods show limitations in fully exploiting the information among the hyperspectral bands. As such, we propose to investigate the chromatic inter-dependencies in their respective hyperspectral embedding space. These embedded features can be fully exploited by querying the inter-channel correlations in a combinatorial manner, with the unique and complementary information efficiently fused into the final prediction. We found such independent modeling and combinatorial excavation mechanisms are extremely beneficial to uncover marginal spectral features, especially in the long wavelength bands. In addition, we have proposed a spatio-spectral attention block and a spectrum-fusion attention module, which greatly facilitates the excavation and fusion of information at both semantically long-range levels and fine-grained pixel levels across all dimensions. Extensive quantitative and qualitative experiments show that our method (dubbed CESST) achieves SOTA performance. Code for this project is at: https://github.com/AlexYangxx/CESST.
We provide a theoretical, numerical, and experimental investigation of the Kerr nonlinearity impact on the performance of a frequency-multiplexed Extreme Learning Machine (ELM). In such ELM, the neuron signals are encoded in the lines of a frequency comb. The Kerr nonlinearity facilitates the randomized neuron connections allowing for efficient information mixing. A programmable spectral filter applies the output weights. The system operates in a continuous-wave regime. Even at low input peak powers, the resulting weak Kerr nonlinearity is sufficient to significantly boost the performance on several tasks. This boost already arises when one uses only the very small Kerr nonlinearity present in a 20-meter long erbium-doped fiber amplifier. In contrast, a subsequent propagation in 540 meters of a single-mode fiber improves the performance only slightly, whereas additional information mixing with a phase modulator does not result in a further improvement at all. We introduce a model to show that, in frequency-multiplexed ELMs, the Kerr nonlinearity mixes information via four-wave mixing, rather than via self- or cross-phase modulation. At low powers, this effect is quartic in the comb-line amplitudes. Numerical simulations validate our experimental results and interpretation.
In the rapidly evolving healthcare industry, platforms now have access to not only traditional medical records, but also diverse data sets encompassing various patient interactions, such as those from healthcare web portals. To address this rich diversity of data, we introduce WellFactor: a method that derives patient profiles by integrating information from these sources. Central to our approach is the utilization of constrained low-rank approximation. WellFactor is optimized to handle the sparsity that is often inherent in healthcare data. Moreover, by incorporating task-specific label information, our method refines the embedding results, offering a more informed perspective on patients. One important feature of WellFactor is its ability to compute embeddings for new, previously unobserved patient data instantaneously, eliminating the need to revisit the entire data set or recomputing the embedding. Comprehensive evaluations on real-world healthcare data demonstrate WellFactor's effectiveness. It produces better results compared to other existing methods in classification performance, yields meaningful clustering of patients, and delivers consistent results in patient similarity searches and predictions.
Automatic speech recognition (ASR) has reached a level of accuracy in recent years, that even outperforms humans in transcribing speech to text. Nevertheless, all current ASR approaches show a certain weakness against ambient noise. To reduce this weakness, audio-visual speech recognition (AVSR) approaches additionally consider visual information from lip movements for transcription. This additional modality increases the computational cost for training models from scratch. We propose an approach, that builds on a pre-trained ASR model and extends it with an adaptive upstream module, that fuses audio and visual information. Since we do not need to train the transformer structure from scratch, our approach requires a fraction of the computational resources compared to traditional AVSR models. Compared to current SOTA systems like AV-HuBERT, our approach achieves an average improvement of 8.3% in word error rate across different model sizes, noise categories and broad SNR range. The approach allows up to 21% smaller models and requires only a fraction of the computational resources for training and inference compared to common AVSR approaches.
Deciphering the oracle bone script plays a significant role in Chinese archaeology and philology. However, it is significantly challenging due to the scarcity of oracle character images. To overcome this issue, we propose Diff-Oracle, based on diffusion models (DMs), to generate sufficient controllable oracle characters. In contrast to most DMs that rely on text prompts, we incorporate a style encoder to control style information during the generation process. This encoder extracts style prompts from existing oracle character images, where style details are converted from a CLIP model into a text embedding format. Inspired by ControlNet, we introduce a content encoder to capture desired content information from content images, ensuring the fidelity of character glyphs. To train Diff-Oracle effectively, we propose to obtain pixel-level paired oracle character images (i.e., style and content images) by a pre-trained image-to-image translation model. Extensive qualitative and quantitative experiments conducted on two benchmark datasets, Oracle-241 and OBC306, demonstrate that our Diff-Oracle outperforms existing generative methods in terms of image generation, further enhancing recognition accuracy. Source codes will be available.