2 November 2022 in Knowledge and Information Systems
14 October 2019 in IEEE Access
“Lack of supervision” is a particularly challenging problem in E-learning or distance learning environments. A wide range of research efforts and technologies have been explored to alleviate its impact by monitoring students' engagement, such as emotion or learning behaviors. However, the current research still lacks multi-dimensional computational measures for analyzing learner's engagement from the interactions that occur in digital learning environment. In this paper, we propose an integrated framework to identify learning engagement from three facets: affect, behavior and cognitive state, which are conveyed by learner's facial expressions, eye movement behaviors and the overall performance during short video learning session. To recognize the three states of learners, three channel data is recorded: 1) video/image sequence captured by camera; 2) eye movement information from a non-intrusive and cost-effective eye tracker; and 3) click stream data from mouse. Based on these modalities, a multi-channel data fusion strategy is designed that concatenates time series features of three channels in the same time segment to predict course learning performance. We also presented a new method to make the self-reported annotations more reliable without using external observers' verification. To validate the approach and methods, 46 participants were invited to attend a representative on-line course that consists of short videos in our designed learning environment. The results demonstrated the effectiveness of the proposed framework and methods in monitoring learning engagement. More importantly, a prototype system was developed to detect learner's emotional and eye behavioral engagement in real-time as well as predict the learning performance of learners after they had completed each short video course.
28 February 2024 in IEEE Transactions on Multimedia
Recently, video recognition is emerging with the help of multi-modal learning, which focuses on integrating distinct modalities to improve the performance or robustness of the model. Although various multi-modal learning methods have been proposed and offer remarkable recognition results, almost all of these methods rely on high-quality manual annotations and assume that modalities among multi-modal data provide semantically relevant information. Unfortunately, the widely used video datasets are usually coarse-annotated or collected from the Internet. Thus, it inevitably contains a portion of noisy labels and noisy correspondence. To address this challenge, we use the audio-visual action recognition task as a proxy and propose a noise-tolerant learning framework to find anti-interference model parameters against both noisy labels and noisy correspondence. Specifically, our method consists of two phases that aim to rectify noise by the inherent correlation between modalities. First, a noise-tolerant contrastive training phase is performed to make the model immune to the possible noisy-labeled data. Despite the benefits brought by contrastive training, it would overfit the noisy correspondence and thus provide false supervision. To alleviate the influence of noisy correspondence, we propose a cross-modal noise estimation component to adjust the consistency between different modalities. As the noisy correspondence existed at the instance level, we further propose a category-level contrastive loss to reduce its interference. Second, in the hybrid-supervised training phase, we calculate the distance metric among features to obtain corrected labels, which are used as complementary supervision to guide the training. Furthermore, due to the lack of suitable datasets, we establish a benchmark of real-world noisy correspondence in audio-visual data by relabeling the Kinetics dataset. Extensive experiments on a wide range of noisy levels demonstrate that our method significantly improves the robustness of the action recognition model and surpasses the baselines by a clear margin.
24 March 2024 in Proceedings of the AAAI Conference on Artificial Intelligence
Generalized Category Discovery (GCD) aims to recognize both known and novel categories in an unlabeled dataset by leveraging another labeled dataset with only known categories. Without considering knowledge transfer from known to novel categories, current methods usually perform poorly on novel categories due to the lack of corresponding supervision. To mitigate this issue, we propose a unified Knowledge Transfer Network (KTN), which solves two obstacles to knowledge transfer in GCD. First, the mixture of known and novel categories in unlabeled data makes it difficult to identify transfer candidates (i.e., samples with novel categories). For this, we propose an entropy-based method that leverages knowledge in the pre-trained classifier to differentiate known and novel categories without requiring extra data or parameters. Second, the lack of prior knowledge of novel categories presents challenges in quantifying semantic relationships between categories to decide the transfer weights. For this, we model different categories with prototypes and treat their similarities as transfer weights to measure the semantic similarities between categories. On the basis of two treatments, we transfer knowledge from known to novel categories by conducting pre-adjustment of logits and post-adjustment of labels for transfer candidates based on the transfer weights between different categories. With the weighted adjustment, KTN can generate more accurate pseudo-labels for unlabeled data, which helps to learn more discriminative features and boost model performance on novel categories. Extensive experiments show that our method outperforms state-of-the-art models on all evaluation metrics across multiple benchmark datasets. Furthermore, different from previous clustering-based methods that can only work offline with abundant data, KTN can be deployed online conveniently with faster inference speed. Code and data are available at https://github.com/yibai-shi/KTN.