AI-synthesized voice technology has the potential to create realistic human voices for beneficial applications, but it can also be misused for malicious purposes. While existing AI-synthesized voice detection models excel in intra-domain evaluation, they face challenges in generalizing across different domains, potentially becoming obsolete as new voice generators emerge. Current solutions use diverse data and advanced machine learning techniques (e.g., domain-invariant representation, self-supervised learning), but are limited by predefined vocoders and sensitivity to factors like background noise and speaker identity. In this work, we introduce an innovative disentanglement framework aimed at extracting domain-agnostic artifact features related to vocoders. Utilizing these features, we enhance model learning in a flat loss landscape, enabling escape from suboptimal solutions and improving generalization. Extensive experiments on benchmarks show our approach outperforms state-of-the-art methods, achieving up to 5.12% improvement in the equal error rate metric in intra-domain and 7.59% in cross-domain evaluations.

Despite significant progress in semi-supervised learning for image object detection, several key issues are yet to be addressed for video object detection: (1) Achieving good performance for supervised video object detection greatly depends on the availability of annotated frames. (2) Despite having large inter-frame correlations in a video, collecting annotations for a large number of frames per video is expensive, time-consuming, and often redundant. (3) Existing semi-supervised techniques on static images can hardly exploit the temporal motion dynamics inherently present in videos. In this paper, we introduce SSVOD, an end-to-end semi-supervised video object detection framework that exploits motion dynamics of videos to utilize large-scale unlabeled frames with sparse annotations. To selectively assemble robust pseudo-labels across groups of frames, we introduce flow-warped predictions from nearby frames for temporal-consistency estimation. In particular, we introduce cross-IoU and cross-divergence based selection methods over a set of estimated predictions to include robust pseudo-labels for bounding boxes and class labels, respectively. To strike a balance between confirmation bias and uncertainty noise in pseudo-labels, we propose confidence threshold based combination of hard and soft pseudo-labels. Our method achieves significant performance improvements over existing methods on ImageNet-VID, Epic-KITCHENS, and YouTube-VIS datasets. Codes are available at https://github.com/enyacgroup/SSVOD.git.

Understanding scenes in movies is crucial for a variety of applications such as video moderation, search, and recommendation. However, labeling individual scenes is a time-consuming process. In contrast, movie level metadata (e.g., genre, synopsis, etc.) regularly gets produced as part of the film production process, and is therefore significantly more commonly available. In this work, we propose a novel contrastive learning approach that uses movie metadata to learn a general-purpose scene representation. Specifically, we use movie metadata to define a measure of movie similarity, and use it during contrastive learning to limit our search for positive scene-pairs to only the movies that are considered similar to each other. Our learned scene representation consistently outperforms existing state-of-the-art methods on a diverse set of tasks evaluated using multiple benchmark datasets. Notably, our learned representation offers an average improvement of 7.9% on the seven classification tasks and 9.7% improvement on the two regression tasks in LVU dataset. Furthermore, using a newly collected movie dataset, we present comparative results of our scene representation on a set of video moderation tasks to demonstrate its generalizability on previously less explored tasks.

Dynamic spectrum access (DSA) enables secondary networks to find and efficiently exploit spectrum opportunities. A key factor to design a DSA network is the spectrum sensing algorithms for multiple channels with multiple users. Multi-user cooperative channel sensing reduces the sensing time, and thus it increases transmission throughput. However, in a multi-channel system, the problem becomes more complex since the benefits of assigning users to sense channels in parallel must also be considered. A sensing schedule, indicating to each user the channel that it should sense at different sensing moments, must be thus created to optimize system performance. In this paper, we formulate the general sensing scheduling optimization problem and then propose several sensing strategies to schedule the users according to network parameters with homogeneous sensors. Later on, we extend the results to heterogeneous sensors and propose a robust scheduling design when we have traffic and channel uncertainty. We propose three sensing strategies, and, within each one of them, several solutions, striking a balance between throughput performance and computational complexity, are proposed. In addition, we show that a sequential channel sensing strategy is the one to be preferred when the sensing time is small, the number of channels is large, and the number of users is small. For all the other cases, a parallel channel sensing strategy is recommended in terms of throughput performance. We also show that a proposed hybrid sequential-parallel channel sensing strategy achieves the best performance in all scenarios at the cost of extra memory and computation complexity.

Dynamic spectrum access (DSA) for secondary networks improves the spectrum utilization by finding spectrum opportunities and exploiting them efficiently. A key factor to design a DSA network is the spectrum sensing algorithms for multiple channels with multiple users. Multi-user cooperative channel sensing reduces the sensing time, thus increasing the transmission throughput. However, in a multi-channel system, the problem becomes more complex since a sensing schedule, indicating to each user the channel that it must sense at different sensing moments, must be created to optimize system performance. In this paper, we first propose a general sensing strategy to schedule the users according to network parameters. We propose three sensing strategies, and within each one of them several solutions striking a balance between throughput performance, memory usage, and computational complexity are proposed. In addition, we show that the proposed sequential sensing strategy is the one to be preferred when the sensing time is small, the number of channels is large, and the number of users is small. For all the other cases, the parallel sensing strategy is recommended in terms of throughput performance. We also show that a proposed hybrid sequential-parallel sensing strategy achieves the best performance in all scenarios at the cost of extra complexity.

This paper focuses on analytical studies of the primary user (PU) traffic classification problem. colorblack{Observing} that the gamma distribution can represent positively skewed data and exponential distribution (popular in communication networks performance analysis literature) it is considered here as the PU traffic descriptor. We investigate two PU traffic classifiers utilizing perfectly measured PU activity (busy) and inactivity (idle) periods: (i) maximum likelihood classifier (MLC) and (ii) multi-hypothesis sequential probability ratio test classifier (MSPRTC). Then, relaxing the assumption on perfect period measurement, we consider a PU traffic observation through channel sampling. For a special case of negligible probability of PU state change in between two samplings, we propose a minimum variance PU busy/idle period length estimator. Later, relaxing the assumption of the complete knowledge of the parameters of the PU period length distribution, we propose two PU traffic classification schemes: (i) estimate-then-classify (ETC), and (ii) average likelihood function (ALF) classifiers considering time domain fluctuation of the PU traffic parameters. Numerical results show that both MLC and MSPRTC are sensitive to the periods measurement errors when the distance among distribution hypotheses is small, and to the distribution parameter estimation errors when the distance among hypotheses is large. For PU traffic parameters with a partial prior knowledge of the distribution, the ETC outperforms ALF when the distance among hypotheses is small, while the opposite holds when the distance is large.

We propose a primary user (PU) traffic distribution classifier for dynamic spectrum access networks based on multi-hypothesis sequential probability ratio test (MSPRT). In specific, we propose two classifiers: (i) an estimate-then-classify classifier, and (ii) a modified MSPRT classifier based on the average likelihood function considering partial knowledge of the PU traffic parameters. Using the sequential algorithm, we show that our proposed classifiers can achieve higher classification performance compared to the traditional maximum likelihood classifier using constant number of samples.

In this work we investigate the bounds on the estimation accuracy of Primary User (PU) traffic parameters with exponentially distributed busy and idle times. We derive closed-form expressions for the Cramér-Rao bounds on the mean squared estimation error for the blind joint estimation of the PU traffic parameters, specifically, the duty cycle, and the mean arrival and departure rates. Moreover, we present the corresponding maximum-likelihood estimators for the traffic parameters and discuss the effect of sensing errors in the joint estimation of PU traffic.

This paper presents a mathematical analysis of the accuracy of estimating Primary User's (PU's) mean duty cycle u, as well as the mean off- and on-times, where the estimation accuracy is expressed in terms of the Cramer-Rao bound on the mean squared estimation error. For estimating u, we derive the mean squared estimation error for uniform, non-uniform, and weighted sample stream averaging, as well as maximum likelihood (ML) estimation. The estimation accuracy of the mean PU off- and on-times is studied when ML estimation is employed. Besides, the impact of spectrum sensing errors on the estimation accuracy is studied analytically for the averaging estimators, while simulation results are used for the ML estimators. Furthermore, we develop algorithms for the blind estimation of the traffic parameters based on the derived theoretical estimation accuracy expressions.

In this paper we present new results on the problem of finding the best channel sensing order for multi-channel Dynamic Spectrum Access (DSA) networks. We start with the general assumption that all Secondary Users (SUs) cooperatively sense each Primary User (PU) channel at one time. Then, the SU sensing results are reported to a DSA base station that schedules SU transmissions in order to maximize DSA network throughput. We then assume that PU traffic parameters are not perfectly known to DSA network and change over time, and propose a novel PU channel sensing order scheme based on the quality of PU traffic estimation. We adopt a maximum likelihood estimator to estimate the traffic statistics of PU channels and derive the Cramer-Rao (CR) bounds for the PU traffic estimation performance. Based on the CR bound and its Gaussian approximation, we analyze the impact of the estimation error on the DSA network throughput by computing a new metric called sensing order confidence, i.e., the probability that the best selected sensing order is not affected by PU traffic estimation errors. Finally, we formulate a convex optimization problem to determine the minimum number of PU channel state samples required for estimating PU traffic parameters after determining a certain constraint on the sensing order confidence metric to achieve the best sensing order.

In order to fully exploit the availability of primary users' (PUs) unused spectrum by secondary users, traffic prediction can be used to increase system throughput. In this paper we propose a PU traffic prediction algorithm based on estimated PU traffic state transition probabilities. The probabilities are obtained via constrained-time PU traffic parameters estimation assuming exponentially distributed PU ON/OFF channel utilization intervals. Moreover, we define prediction regions for the estimated parameters where optimal traffic prediction is possible. Finally, we theoretically quantify the prediction confidence as a function of the prediction time, the total estimation time period and the number of samples used for estimation.