Given a query in one modality, video-text retrieval aims to retrieve the most similar samples from the database in another modality. The primary challenge lies in the alignment of fine-grained topology, including objects and interactions among diverse objects. In this paper, we introduce a fine-grained relationship alignment network. Specifically, we adaptively recalibrate frame-wise features of videos and extract fine-grained relationship features, including semantic objects and structural interactions among various objects. Correspondingly, we parse texts into dual paths and encode semantic and structural features. Finally, we combine semantic and structural features to align videos and texts. Remarkably, a negative sample enhanced ranking mechanism is proposed to optimize the network. Experiments on public datasets demonstrate the advantages of our model.

Multi-modal semantic understanding requires integrating information from different modalities to extract users’ real intention behind words. Most previous work applies a dual-encoder structure to separately encode image and text, but fails to learn cross-modal feature alignment, making it hard to achieve cross-modal deep information interaction. This paper proposes a novel CLIP-guided contrastive-learning-based architecture to perform multi-modal feature alignment, which projects the features derived from different modalities into a unified deep space. On multi-modal sarcasm detection (MMSD) and multi-modal sentiment analysis (MMSA) tasks, the experimental results show that our proposed model significantly outperforms several baselines, and our feature alignment strategy brings obvious performance gain over models with different aggregating methods and models even enriched with knowledge. More importantly, our model is simple to implement without using task-specific external knowledge, and thus can easily migrate to other multi-modal tasks. Our source codes are available at https://github.com/ChangKe123/CLFA.

This brief presents a higher-order vector DEM for multibit discrete-time (DT)$\boldsymbol \Sigma \boldsymbol \Delta $modulators ($\boldsymbol \Sigma \boldsymbol \Delta $Ms) to achieve higher linearity. By using the proposed vector filter (VF) with a poles-splitting technique, the root locus outside the unit circle of higher-order DEM can be eliminated, leading to the DAC mismatch-shaping stability will not be limited by the closed-loop poles of the vector DEM, thus a self-stabilizing vector DEM at higher order can be achieved. Besides, a complexity-reduced vector quantizer (VQ) using the segmented architecture is proposed. Compared to recent works, the proposed technique can achieve higher-order mismatch shaping and decouple the correlation between the mismatch errors and the input pattern, leading to stronger suppression of DAC mismatch and canceling in-band tones.

Real-time bidding (RTB) based display advertising has become one of the key technological advances in computational advertising. RTB enables advertisers to buy individual ad impressions via an auction in real-time and facilitates the evaluation and the bidding of individual impressions across multiple advertisers. In RTB, the advertisers face three main challenges when optimizing their bidding strategies, namely (i) estimating the utility (e.g., conversions, clicks) of the ad impression, (ii) forecasting the market value (thus the cost) of the given ad impression, and (iii) deciding the optimal bid for the given auction based on the first two. Previous solutions assume the first two are solved before addressing the bid optimization problem. However, these challenges are strongly correlated and dealing with any individual problem independently may not be globally optimal. In this paper, we propose Bidding Machine, a comprehensive learning to bid framework, which consists of three optimizers dealing with each challenge above, and as a whole, jointly optimizes these three parts. We show that such a joint optimization would largely increase the campaign effectiveness and the profit. From the learning perspective, we show that the bidding machine can be updated smoothly with both offline periodical batch or online sequential training schemes. Our extensive offline empirical study and online A/B testing verify the high effectiveness of the proposed bidding machine.