Existing video hashing techniques for large-scale video retrieval often overlook inherent audio signals, which can potentially compromise retrieval performance. Incorporating both visual and audio signals, however, complicates neural architecture design, rendering the manual crafting of joint audio-visual neural network models challenging. To address this issue, we propose AV-NAS, a method that leverages data-driven Neural Architecture Search (NAS) within a tailored audio-visual network space to automatically discover the optimal video hashing network. Our approach offers: (1) a versatile multi-level semantic architecture based on audio-visual signals, defining a mixed search space encompassing diverse network modules such as MLP, CNN, Transformer, and Mamba, as well as operations like Add, Hadamard, SiLU, LayerNorm, and Skip; (2) a differentiable relaxation of the combinatorial search problem, converting it into a unified differentiable optimization problem which we tackle through our ''coarse search-pruning-finetuning'' strategy. Our experiments on large-scale video datasets show that AV-NAS can discover architectures distinct from expert designs and lead to substantial performance improvements over current state-of-the-art methods including the recently emerged AVHash.

Deep hashing presents an effective strategy for large-scale image retrieval. Current hashing methods are generally categorized by their supervision types: point-wise, pairwise, and list-wise. Recent advancements in point-wise methods (e.g., CSQ, MDS) have significantly enhanced retrieval performance across diverse datasets by pre-assigning a hash center to each class, thereby improving the discriminability of the resultant hash codes. However, these methods employ purely data-independent algorithms for generating hash centers, overlooking the semantic connections between different classes, which, we argue, could degrade retrieval performance. To tackle this problem, this article expands on the newly emerged concept of “hash centers” to introduce “ semantic hash centers,” which posits that hash centers of semantically related classes should exhibit closer Hamming distances, while those of unrelated classes should be more distant. Based on this hypothesis, we propose a three-stage framework, termed Semantic Hash Centers (SHC), to produce hash codes that preserve semantics. First, we build a classification network to detect semantic similarities between classes, and utilize a data-dependent approach to similarity calculation that can adapt to varied data distributions. Next, we develop a new optimization algorithm to generate SHC. This algorithm not only maintains semantic relatedness among hash centers but also integrates a constraint to ensure a minimum distance between them, addressing the issue of excessively proximate hash centers potentially impairing retrieval performance. Finally, we train a deep hashing network with the above generated SHC to convert each image into a binary hash code. Experiments on large-scale image retrieval across several public datasets demonstrate that SHC generates more discriminative hash codes, markedly enhancing retrieval performance. Specifically, in terms of the mAP@100, mAP@1000, and mAP@ALL metrics, SHC records average improvements of +6.24%, +6.68%, and +10.39%, respectively, over the most competitive existing methods. The code of our SHC project is available at https://github.com/cc752424640/Deep-Hashing-with-Semantic-Hash-Centers-for-Image-Retrieval .

Prompt learning has been proven to be quite an effective technique for adapting large visual-language models (LVLMs) to downstream tasks via few-shot learning. Early methods often rely on a single prompt, which is insufficient for comprehensively representing a class. Subsequent efforts have explored multiple prompts to further enhance the adaptability and performance of LVLMs. However, these methods primarily focus on learning a set of more discriminative prompts, overlooking their generalizability. To learn prompts that are more balanced in both generalization and discrimination, we propose a novel multi-prompt learning approach, Masked Multi-Prompt Learning with Knowledge Mixing (dubbed TriMPL), which contains two pivotal mechanisms: (1) knowledge mixing to enhance the generalization of each individual prompt and (2) prompt masking to boost the prompt set's overall robustness. With respect to knowledge mixing, it progressively injects the general knowledge of handcrafted prompts into each learnable prompt at different Transformer encoding stages. While for prompt masking, of which the critical insight is that an optimal set of prompts should exhibit independence, allowing accurate predictions with just a subset of prompts. During training, TriMPL randomly masks some prompts to enhance the overall robustness of the learned prompts for image classification. We evaluate the effectiveness of TriMPL under three settings: (1) base-to-new generalization, (2) cross-dataset transfer, and (3) domain generalization. Extensive experiments demonstrate that TriMPL is capable of learning a set of effective prompts, achieving superior performance to quite a few state-of-the-art competitors.

This paper proposes a novel method, Enhanced Discrete Multi-modal Hashing (EDMH), which learns binary codes and hash functions simultaneously from the pairwise similarity matrix of data for large-scale cross-view retrieval. EDMH distinguishes itself from existing methods by considering not just the binarization constraint but also the balance and decorrelation constraints. Although those additional discrete constraints make the optimization problem of EDMH look a lot more complicated, we are actually able to develop a fast iterative learning algorithm in the alternating optimization framework for it, as after introducing a couple of auxiliary variables each subproblem of optimization turns out to have closed-form solutions. It has been confirmed by extensive experiments that EDMH can consistently deliver better retrieval performances than state-of-the-art MH methods at lower computational costs.

Due to the exponential growth of multimedia data, multi-modal hashing as a promising technique to make cross-view retrieval scalable is attracting more and more attention. However, most of the existing multi-modal hashing methods either divide the learning process unnaturally into two separate stages or treat the discrete optimization problem simplistically as a continuous one, which leads to suboptimal results. Recently, a few discrete multi-modal hashing methods that try to address such issues have emerged, but they still ignore several important discrete constraints (such as the balance and decorrelation of hash bits). In this paper, we overcome those limitations by proposing a novel method named “Enhanced Discrete Multi-modal Hashing (EDMH)” which learns binary codes and hashing functions simultaneously from the pairwise similarity matrix of data, under the aforementioned discrete constraints. Although the model of EDMH looks a lot more complex than the other models for multi-modal hashing, we are actually able to develop a fast iterative learning algorithm for it, since the subproblems of its optimization all have closed-form solutions after introducing a couple of auxiliary variables. Our experimental results on three real-world datasets have revealed the usefulness of those previously ignored discrete constraints and demonstrated that EDMH not only performs much better than state-of-the-art competitors according to several retrieval metrics but also runs much faster than most of them.

Hashing, which represents data items as compact binary codes, has been becoming a more and more popular technique, e.g., for large-scale image retrieval, owing to its super fast search speed as well as its extremely economical memory consumption. However, existing hashing methods all try to learn binary codes from artificially balanced datasets which are not commonly available in real-world scenarios. In this paper, we propose Long-Tail Hashing Network (LTHNet), a novel two-stage deep hashing approach that addresses the problem of learning to hash for more realistic datasets where the data labels roughly exhibit a long-tail distribution. Specifically, the first stage is to learn relaxed embeddings of the given dataset with its long-tail characteristic taken into account via an end-to-end deep neural network; the second stage is to binarize those obtained embeddings. A critical part of LTHNet is its dynamic meta-embedding module extended with a determinantal point process which can adaptively realize visual knowledge transfer between head and tail classes, and thus enrich image representations for hashing. Our experiments have shown that LTHNet achieves dramatic performance improvements over all state-of-the-art competitors on long-tail datasets, with no or little sacrifice on balanced datasets. Further analyses reveal that while to our surprise directly manipulating class weights in the loss function has little effect, the extended dynamic meta-embedding module, the usage of cross-entropy loss instead of square loss, and the relatively small batch-size for training all contribute to LTHNet's success.

This paper explores technological trajectory transition in the perspective of innovation ecosystem and their effect on innovation performance of latecomers in market. A structural equation model is developed and tested with data collected from 366 firms in China. In specific, this paper categories technological trajectory transition creative accumulative technological trajectory transition (CCT) and creative disruptive technological trajectory transition (CDT). The results indicate that firms' organizational learning ability positively affect their technological trajectory transition and innovation performance. Firms' network relationship strength negatively affects their technological trajectory transition and positively affect their innovation performance. Governments' environmental concerns positively affect firms' technological trajectory transition and their innovation performance, whereas firms' environmental concerns do not. CCT does not positively affect their innovation performance. In contrast, CDT positively affects their innovation performance.

This paper identifies three stages in the radical technological innovation process, namely formation process in niches, breaking out of niches and entering regimes, and new regime formation. It then adopts Multi-level Perspective (MLP) to explore the formation process, operating mechanism, breakthrough path, and impact factors of radical technological innovation. A three-phase model, which includes formation of radical innovation, breakout of radical innovation, and new regimes construction, is proposed to analyze radical technological innovation. The model is adopted in a case study to analyze the leapfrogging development of technologies in China’s mobile communication industry. This paper enriches technological innovation theory and provides supports for policy making and guidance for industries/enterprises practices regarding technological innovation in emerging economies.

The paper explores the relationship among technological innovation, technological trajectory transition, and firms’ innovation performance. Technological innovation is studied from the perspectives of innovation novelty and innovation openness. Technological trajectory transition is categorized into creative cumulative technological trajectory transition and creative disruptive technological trajectory transition. A structural equation model is developed and tested with data collected by surveying 366 Chinese firms. The results indicate that both innovation novelty and innovation openness positively affects creative cumulative technological trajectory transition as well as creative disruptive technological trajectory transition. Innovation openness and creative disruptive technological trajectory transition both positively affect firms’ innovation performance. However, neither innovation novelty nor creative cumulative technological trajectory transition positively affects firms’ innovation performance. Implications for managers and directions for future studies are discussed.

Lifelong topic model (LTM), an emerging paradigm for never-ending topic learning, aims to yield higher-quality topics as time passes through knowledge accumulated from the past yet learned for the future. In this paper, we propose a novel lifelong topic model based on non-negative matrix factorization (NMF), called Affinity Regularized NMF for LTM (NMF-LTM), which to our best knowledge is distinctive from the popular LDA-based LTMs. NMF-LTM achieves lifelong learning by introducing word-word graph Laplacian as semantic affinity regularization. Other priors such as sparsity, diversity, and between-class affinity are incorporated as well for better performance, and a theoretical guarantee is provided for the algorithmic convergence to a local minimum. Extensive experiments on various public corpora demonstrate the effectiveness of NMF-LTM, particularly its human-like behaviors in two carefully designed learning tasks and the ability in topic modeling of big data. A further exploration of semantic relatedness in knowledge graphs and a case study on a large-scale real-world corpus exhibit the strength of NMF-LTM in discovering high-quality topics in an efficient and robust way.

Dynamic topic models (DTM) are of great use toanalyze the evolution of unobserved topics of a text collectionover time. Recent years have witnessed the explosive growth ofstreaming text data emerging from online media, which createsan unprecedented need for DTMs for timely event analysis. While there have been some matrix factorization methods inthe literature for dynamic topic modeling, further study is stillin great need to model emerging, evolving and fading topicsin a more natural and effective way. In light of this, we firstpropose a matrix factorization model called SONMFSR (SoftOrthogonal NMF with Sparse Representation), which makes fulluse of soft orthogonal and sparsity constraints for static topicmodeling. Furthermore, by introducing the constraints of emerging, evolving and fading topics to SONMFSR, we easily obtain a novel DTM called SONMFSRd for dynamic event analysis. Extensive experiments on two public corpora demonstrate the superiority of SONMFSRd to some state-of-the-art DTMs in both topic detection and tracking. In particular, SONMFSRd shows great potential in real-world applications, where popular topics in Two Sessions 2015 are captured and traced dynamically for possible insights.