arXiv subjects

Wen Wu

Publications and source records attributed to Wen Wu.

At least 145 records · Page 8Linked to original sources

arXiv2022

Distribution-based Emotion Recognition in Conversation

Automatic emotion recognition in conversation (ERC) is crucial for emotion-aware conversational artificial intelligence. This paper proposes a distribution-based framework that formulates ERC as a sequence-to-sequence problem for emotion distribution estimation. The inherent ambiguity of emotions and the subjectivity of human perception lead to disagreements in emotion labels, which is handled naturally in our framework from the perspective of uncertainty estimation in emotion distributions. A Bayesian training loss is introduced to improve the uncertainty estimation by conditioning each emotional state on an utterance-specific Dirichlet prior distribution. Experimental results on the IEMOCAP dataset show that ERC outperformed the single-utterance-based system, and the proposed distribution-based ERC methods have not only better classification accuracy, but also show improved uncertainty estimation.

cs.CL↗

arXiv2022

DACSR: Decoupled-Aggregated End-to-End Calibrated Sequential Recommendation

Sequential recommendations have made great strides in accurately predicting the future behavior of users. However, seeking accuracy alone may bring side effects such as unfair and overspecialized recommendation results. In this work, we focus on the calibrated recommendations for sequential recommendation, which is connected to both fairness and diversity. On the one hand, it aims to provide fairer recommendations whose preference distributions are consistent with users' historical behaviors. On the other hand, it can improve the diversity of recommendations to a certain degree. But existing methods for calibration have mainly relied on the post-processing on the candidate lists, which require more computation time in generating recommendations. In addition, they fail to establish the relationship between accuracy and calibration, leading to the limitation of accuracy. To handle these problems, we propose an end-to-end framework to provide both accurate and calibrated recommendations for sequential recommendation. We design an objective function to calibrate the interests between recommendation lists and historical behaviors. We also provide distribution modification approaches to improve the diversity and mitigate the effect of imbalanced interests. In addition, we design a decoupled-aggregated model to improve the recommendation. The framework assigns two objectives to two individual sequence encoders, and aggregates the outputs by extracting useful information. Experiments on benchmark datasets validate the effectiveness of our proposed model.

cs.IR↗

arXiv2022

Personalized QoE Enhancement for Adaptive Video Streaming: A Digital Twin-Assisted Scheme

In this paper, we present a digital twin (DT)-assisted adaptive video streaming scheme to enhance personalized quality-of-experience (PQoE). Since PQoE models are user-specific and time-varying, existing schemes based on universal and time-invariant PQoE models may suffer from performance degradation. To address this issue, we first propose a DT-assisted PQoE model construction method to obtain accurate user-specific PQoE models. Specifically, user DTs (UDTs) are respectively constructed for individual users, which can acquire and utilize users' data to accurately tune PQoE model parameters in real time. Next, given the obtained PQoE models, we formulate a resource management problem to maximize the overall long-term PQoE by taking the dynamics of user' locations, video content requests, and buffer statuses into account. To solve this problem, a deep reinforcement learning algorithm is developed to jointly determine segment version selection, and communication and computing resource allocation. Simulation results on the real-world dataset demonstrate that the proposed scheme can effectively enhance PQoE compared with benchmark schemes.

eess.IV↗

arXiv2022

Climate and Weather: Inspecting Depression Detection via Emotion Recognition

Automatic depression detection has attracted increasing amount of attention but remains a challenging task. Psychological research suggests that depressive mood is closely related with emotion expression and perception, which motivates the investigation of whether knowledge of emotion recognition can be transferred for depression detection. This paper uses pretrained features extracted from the emotion recognition model for depression detection, further fuses emotion modality with audio and text to form multimodal depression detection. The proposed emotion transfer improves depression detection performance on DAIC-WOZ as well as increases the training stability. The analysis of how the emotion expressed by depressed individuals is further perceived provides clues for further understanding of the relationship between depression and emotion.

cs.CL↗

arXiv2022

Long-Tail Session-based Recommendation from Calibration

Accurate predictions in session-based recommendations have progressed, but a few studies have focused on skewed recommendation lists caused by popularity bias. Existing models for mitigating popularity bias have attempted to reduce the overconcentration of popular items by amplifying scores of less popular items. However, they normally ignore the users' different preferences toward long-tail items. Thus, we incorporate calibration, where calibrated recommendations reflect the users' interests in recommendation lists with appropriate proportions, to mitigate the popularity bias from the user's perspective. Specifically, we propose a calibration module to predict the ratio of tail items in the recommendation list from the session representation, and align it to the ongoing session. Additionally, we utilize a two-stage curriculum training strategy to improve prediction in the calibration module. Experiments on benchmark datasets show that our model can both achieve the competitive accuracy of recommendation and provide more tail items.

cs.IR↗

arXiv2022

Multi-channel Attentive Graph Convolutional Network With Sentiment Fusion For Multimodal Sentiment Analysis

Nowadays, with the explosive growth of multimodal reviews on social media platforms, multimodal sentiment analysis has recently gained popularity because of its high relevance to these social media posts. Although most previous studies design various fusion frameworks for learning an interactive representation of multiple modalities, they fail to incorporate sentimental knowledge into inter-modality learning. This paper proposes a Multi-channel Attentive Graph Convolutional Network (MAGCN), consisting of two main components: cross-modality interactive learning and sentimental feature fusion. For cross-modality interactive learning, we exploit the self-attention mechanism combined with densely connected graph convolutional networks to learn inter-modality dynamics. For sentimental feature fusion, we utilize multi-head self-attention to merge sentimental knowledge into inter-modality feature representations. Extensive experiments are conducted on three widely-used datasets. The experimental results demonstrate that the proposed model achieves competitive performance on accuracy and F1 scores compared to several state-of-the-art approaches.

cs.CL↗

arXiv2021

On the 2-binomial complexity of the generalized Thue-Morse words

In this paper, we study the $2$-binomial complexity $b_{\mathbf{t}_{m},2}(n)$ of the generalized Thue-Morse words $\mathbf{t}_{m}$ for every integer $m\geq 3$. We obtain the exact value of $b_{\mathbf{t}_{m},2}(n)$ for every integer $n\geq m^{2}$. As a consequence, $b_{\mathbf{t}_{m},2}(n)$ is ultimately periodic with period $m^{2}$. This result partially answers a question of M. Lejeune, J. Leroy and M. Rigo [Computing the $k$-binomial complexity of the Thue-Morse word, J. Comb. Theory Ser. A, {\bf 176} (2020) 105284].

math.CO↗

arXiv2021

AI-Native Network Slicing for 6G Networks

With the global roll-out of the fifth generation (5G) networks, it is necessary to look beyond 5G and envision the 6G networks. The 6G networks are expected to have space-air-ground integrated networks, advanced network virtualization, and ubiquitous intelligence. This article presents an artificial intelligence (AI)-native network slicing architecture for 6G networks to enable the synergy of AI and network slicing, thereby facilitating intelligent network management and supporting emerging AI services. AI-based solutions are first discussed across network slicing lifecycle to intelligently manage network slices, i.e., AI for slicing. Then, network slicing solutions are studied to support emerging AI services by constructing AI instances and performing efficient resource management, i.e., slicing for AI. Finally, a case study is presented, followed by a discussion of open research issues that are essential for AI-native network slicing in 6G networks.

cs.NI↗

arXiv2021

Emotion recognition by fusing time synchronous and time asynchronous representations

In this paper, a novel two-branch neural network model structure is proposed for multimodal emotion recognition, which consists of a time synchronous branch (TSB) and a time asynchronous branch (TAB). To capture correlations between each word and its acoustic realisation, the TSB combines speech and text modalities at each input window frame and then does pooling across time to form a single embedding vector. The TAB, by contrast, provides cross-utterance information by integrating sentence text embeddings from a number of context utterances into another embedding vector. The final emotion classification uses both the TSB and the TAB embeddings. Experimental results on the IEMOCAP dataset demonstrate that the two-branch structure achieves state-of-the-art results in 4-way classification with all common test setups. When using automatic speech recognition (ASR) output instead of manually transcribed reference text, it is shown that the cross-utterance information considerably improves the robustness against ASR errors. Furthermore, by incorporating an extra class for all the other emotions, the final 5-way classification system with ASR hypotheses can be viewed as a prototype for more realistic emotion recognition systems.

cs.CL↗

arXiv2021

Criteria for apwenian sequences

In 1998, Allouche, Peyrière, Wen and Wen showed that the Hankel determinant $H_n$ of the Thue-Morse sequence over $\{-1,1\}$ satisfies $H_n/2^{n-1}\equiv 1~(\mathrm{mod}~2)$ for all $n\geq 1$. Inspired by this result, Fu and Han introduced \emph{apwenian} sequences over $\{-1,1\}$, namely, $\pm 1$ sequences whose Hankel determinants satisfy $H_n/2^{n-1}\equiv 1~(\mathrm{mod}~2)$ for all $n\geq 1$, and proved with computer assistance that a few sequences are apwenian. In this paper, we obtain an easy to check criterion for apwenian sequences, which allows us to determine all apwenian sequences that are fixed points of substitutions of constant length. Let $f(z)$ be the generating functions of such apwenian sequences. We show that for all integer $b\ge 2$ with $f(1/b)\neq 0$, the real number $f(1/b)$ is transcendental and its irrationality exponent is equal to $2$. Besides, we also derive a criterion for zero-one apwenian sequences whose Hankel determinants satisfy $H_n\equiv 1~(\mathrm{mod}~2)$ for all $n\geq 1$. We find that the only zero-one apwenian sequence, among all fixed points of substitutions of constant length, is the period-doubling sequence. Various examples of apwenian sequences given by substitutions with projection are also given. Furthermore, we prove that all Sturmian sequences over $\{-1,1\}$ or $\{0,1\}$ are not apwenian. And we conjecture that fixed points of substitution of non-constant length over $\{-1,1\}$ or $\{0,1\}$ can not be apwenian.

math.NT↗

arXiv2020

Dynamic RAN Slicing for Service-Oriented Vehicular Networks via Constrained Learning

In this paper, we investigate a radio access network (RAN) slicing problem for Internet of vehicles (IoV) services with different quality of service (QoS) requirements, in which multiple logically-isolated slices are constructed on a common roadside network infrastructure. A dynamic RAN slicing framework is presented to dynamically allocate radio spectrum and computing resource, and distribute computation workloads for the slices. To obtain an optimal RAN slicing policy for accommodating the spatial-temporal dynamics of vehicle traffic density, we first formulate a constrained RAN slicing problem with the objective to minimize long-term system cost. This problem cannot be directly solved by traditional reinforcement learning (RL) algorithms due to complicated coupled constraints among decisions. Therefore, we decouple the problem into a resource allocation subproblem and a workload distribution subproblem, and propose a two-layer constrained RL algorithm, named Resource Allocation and Workload diStribution (RAWS) to solve them. Specifically, an outer layer first makes the resource allocation decision via an RL algorithm, and then an inner layer makes the workload distribution decision via an optimization subroutine. Extensive trace-driven simulations show that the RAWS effectively reduces the system cost while satisfying QoS requirements with a high probability, as compared with benchmarks.

cs.LG↗

arXiv2020

Self-similarity of $\mathcal{P}$-positions of $(2n+1)$-dimensional Wythoff's game

Wythoff's game as a classic combinatorial game has been well studied. In this paper, we focus on $(2n+1)$-dimensional Wythoff's game; that is the Wythoff's game with $(2n+1)$ heaps. We characterize their $\mathcal{P}$-positions explicitly and show that they have self-similar structures. In particular, the set of all $\mathcal{P}$-positions of $3$-dimensional Wythoff's game generates the well-known fractal set---the Sierpinski sponge.

math.CO↗

arXiv2020

Deep Reinforcement Learning for Delay-Oriented IoT Task Scheduling in Space-Air-Ground Integrated Network

In this paper, we investigate a computing task scheduling problem in space-air-ground integrated network (SAGIN) for delay-oriented Internet of Things (IoT) services. In the considered scenario, an unmanned aerial vehicle (UAV) collects computing tasks from IoT devices and then makes online offloading decisions, in which the tasks can be processed at the UAV or offloaded to the nearby base station or the remote satellite. Our objective is to design a task scheduling policy that minimizes offloading and computing delay of all tasks given the UAV energy capacity constraint. To this end, we first formulate the online scheduling problem as an energy-constrained Markov decision process (MDP). Then, considering the task arrival dynamics, we develop a novel deep risk-sensitive reinforcement learning algorithm. Specifically, the algorithm evaluates the risk, which measures the energy consumption that exceeds the constraint, for each state and searches the optimal parameter weighing the minimization of delay and risk while learning the optimal policy. Extensive simulation results demonstrate that the proposed algorithm can reduce the task processing delay by up to 30% compared to probabilistic configuration methods while satisfying the UAV energy capacity constraint.

cs.LG↗

arXiv2020

A Mathematical Framework for Estimating Risk of Airborne Transmission of COVID-19 with Application to Face Mask Use and Social Distancing

A mathematical model for estimating the risk of airborne transmission of a respiratory infection such as COVID-19, is presented. The model employs basic concepts from fluid dynamics and incorporates the known scope of factors involved in the airborne transmission of such diseases. Simplicity in the mathematical form of the model is by design, so that it can serve not only as a common basis for scientific inquiry across disciplinary boundaries, but also be understandable by a broad audience outside science and academia. The caveats and limitations of the model are discussed in detail. The model is used to assess the protection from transmission afforded by face coverings made from a variety of fabrics. The reduction in transmission risk associated with increased physical distance between the host and susceptible is also quantified by coupling the model with available data on scalar dispersion in canonical flows. Finally, the effect of the level of physical activity (or exercise intensity) of the host and the susceptible in enhancing transmission risk, is also assessed.

q-bio.PE↗

arXiv2020

Hankel determinants of a Sturmian sequence

Let $τ$ be the substitution $1\to 101$ and $0\to 1$ on the alphabet $\{0,1\}$. The fixed point of $τ$ leading by 1, denoted by $\mathbf{s}$, is a Sturmian sequence. We first give a characterization of $\mathbf{s}$ using $f$-representation. Then we show that the distribution of zeros in the determinants induces a partition of integer lattices in the first quadrant. Combining those properties, we give the explicit values of the Hankel determinants $H_{m,n}$ of $\mathbf{s}$ for all $m\ge 0$ and $n\ge 1$.

math.CO↗

arXiv2020

Efficient Hybrid Beamforming with Anti-Blockage Design for High-Speed Railway Communications

Future railway is expected to accommodate both train operation services and passenger broadband services. The millimeter wave (mmWave) communication is a promising technology in providing multi-gigabit data rates to onboard users. However, mmWave communications suffer from severe propagation attenuation and vulnerability to blockage, which can be very challenging in high-speed railway (HSR) scenarios. In this paper, we investigate efficient hybrid beamforming (HBF) design for train-to-ground communications. First, we develop a two-stage HBF algorithm in blockage-free scenarios. In the first stage, the minimum mean square error method is adopted for optimal hybrid beamformer design with low complexity and fast convergence; in the second stage, the orthogonal matching pursuit method is utilized to approximately recover the analog and digital beamformers. Second, in blocked scenarios, we design an anti-blockage scheme by adaptively invoking the proposed HBF algorithm, which can efficiently deal with random blockages. Extensive simulation results are presented to show the sum rate performance of the proposed algorithms under various configurations, including transmission power, velocity of the train, blockage probability, etc. It is demonstrated that the proposed anti-blockage algorithm can improve the effective rate by 20% in severely-blocked scenarios while maintaining low outage probability.

cs.IT↗

arXiv2020

Pyramid Focusing Network for mutation prediction and classification in CT images

Predicting the mutation status of genes in tumors is of great clinical significance. Recent studies have suggested that certain mutations may be noninvasively predicted by studying image features of the tumors from Computed Tomography (CT) data. Currently, this kind of image feature identification method mainly relies on manual processing to extract generalized image features alone or machine processing without considering the morphological differences of the tumor itself, which makes it difficult to achieve further breakthroughs. In this paper, we propose a pyramid focusing network (PFNet) for mutation prediction and classification based on CT images. Firstly, we use Space Pyramid Pooling to collect semantic cues in feature maps from multiple scales according to the observation that the shape and size of the tumors are varied.Secondly, we improve the loss function based on the consideration that the features required for proper mutation detection are often not obvious in cross-sections of tumor edges, which raises more attention to these hard examples in the network. Finally, we devise a training scheme based on data augmentation to enhance the generalization ability of networks. Extensively verified on clinical gastric CT datasets of 20 testing volumes with 63648 CT images, our method achieves the accuracy of 94.90% in predicting the HER-2 genes mutation status of at the CT image.

cs.CV↗

arXiv2020

Stieltjes continued fractions related to the Paperfolding sequence and Rudin-Shapiro sequence

We investigate two Stieltjes continued fractions given by the paperfolding sequence and the Rudin-Shapiro sequence. By explicitly describing certain subsequences of the convergents $P_n(x)/Q_n(x)$ modulo $4$, we give the formal power series expansions (modulo $4$) of these two continued fractions and prove that they are congruent modulo $4$ to algebraic series in $\mathbb{Z}[[x]]$. Therefore, the coefficient sequences of the formal power series expansions are $2$-automatic. Write $Q_{n}(x)=\sum_{i\ge 0}a_{n,i}x^{i}$. Then $(Q_{n}(x))_{n\ge 0}$ defines a two-dimensional coefficient sequence $(a_{n,i})_{n,i\ge 0}$. We prove that the coefficient sequences $(a_{n,i}\mod 4)_{n\ge 0}$ introduced by both $(Q_{n}(x))_{n\ge 0}$ and $(P_{n}(x))_{n\ge 0}$ are $2$-automatic for all $i\ge 0$. Moreover, the pictures of these two dimensional coefficient sequences modulo $4$ present a kind of self-similar phenomenon.

math.NT↗