Click-Through Rate (CTR) prediction holds a pivotal place in online advertising and recommender systems since CTR prediction performance directly influences the overall satisfaction of the users and the revenue generated by companies. Even so, CTR prediction is still an active area of research since it involves accurately modelling the preferences of users based on sparse and high-dimensional features where the higher-order interactions of multiple features can lead to different outcomes. Most CTR prediction models have relied on a single fusion and interaction learning strategy. The few CTR prediction models that have utilized multiple interaction modelling strategies have treated each interaction to be self-contained. In this paper, we propose a novel model named STEC that reaps the benefits of multiple interaction learning approaches in a single unified architecture. Additionally, our model introduces residual connections from different orders of interactions which boosts the performance by allowing lower level interactions to directly affect the predictions. Through extensive experiments on four real-world datasets, we demonstrate that STEC outperforms existing state-of-the-art approaches for CTR prediction thanks to its greater expressive capabilities.
With the increasing complexity and scale of click-through rate (CTR) prediction tasks in online advertising and recommendation systems, accurately estimating the importance of features has become a critical aspect of developing effective models. In this paper, we propose an attention-based approach that leverages max and mean pooling operations, along with a bit-wise attention mechanism, to enhance feature importance estimation in CTR prediction. Traditionally, pooling operations such as max and mean pooling have been widely used to extract relevant information from features. However, these operations can lead to information loss and hinder the accurate determination of feature importance. To address this challenge, we propose a novel attention architecture that utilizes a bit-based attention structure that emphasizes the relationships between all bits in features, together with maximum and mean pooling. By considering the fine-grained interactions at the bit level, our method aims to capture intricate patterns and dependencies that might be overlooked by traditional pooling operations. To examine the effectiveness of the proposed method, experiments have been conducted on three public datasets. The experiments demonstrated that the proposed method significantly improves the performance of the base models to achieve state-of-the-art results.
The classification loss functions used in deep neural network classifiers can be grouped into two categories based on maximizing the margin in either Euclidean or angular spaces. Euclidean distances between sample vectors are used during classification for the methods maximizing the margin in Euclidean spaces whereas the Cosine similarity distance is used during the testing stage for the methods maximizing margin in the angular spaces. This paper introduces a novel classification loss that maximizes the margin in both the Euclidean and angular spaces at the same time. This way, the Euclidean and Cosine distances will produce similar and consistent results and complement each other, which will in turn improve the accuracies. The proposed loss function enforces the samples of classes to cluster around the centers that represent them. The centers approximating classes are chosen from the boundary of a hypersphere, and the pairwise distances between class centers are always equivalent. This restriction corresponds to choosing centers from the vertices of a regular simplex. There is not any hyperparameter that must be set by the user in the proposed loss function, therefore the use of the proposed method is extremely easy for classical classification problems. Moreover, since the class samples are compactly clustered around their corresponding means, the proposed classifier is also very suitable for open set recognition problems where test samples can come from the unknown classes that are not seen in the training phase. Experimental studies show that the proposed method achieves the state-of-the-art accuracies on open set recognition despite its simplicity.
Robust object tracking requires knowledge of tracked objects' appearance, motion and their evolution over time. Although motion provides distinctive and complementary information especially for fast moving objects, most of the recent tracking architectures primarily focus on the objects' appearance information. In this paper, we propose a two-stream deep neural network tracker that uses both spatial and temporal features. Our architecture is developed over ATOM tracker and contains two backbones: (i) 2D-CNN network to capture appearance features and (ii) 3D-CNN network to capture motion features. The features returned by the two networks are then fused with attention based Feature Aggregation Module (FAM). Since the whole architecture is unified, it can be trained end-to-end. The experimental results show that the proposed tracker TRAT (TRacking by ATtention) achieves state-of-the-art performance on most of the benchmarks and it significantly outperforms the baseline ATOM tracker.