Searcharxiv⌕ Search

arXiv subjects

Mingjie Chen

Publications and source records attributed to Mingjie Chen.

26 records · Page 2Linked to original sources

One-off Negative Sequential Pattern Mining

Negative sequential pattern mining (SPM) is an important SPM research topic. Unlike positive SPM, negative SPM can discover events that should have occurred but have not occurred, and it can be used for financial risk management and fraud detection. However, existing methods generally ignore the repetitions of the pattern and do not consider gap constraints, which can lead to mining results containing a large number of patterns that users are not interested in. To solve this problem, this paper discovers frequent one-off negative sequential patterns (ONPs). This problem has the following two characteristics. First, the support is calculated under the one-off condition, which means that any character in the sequence can only be used once at most. Second, the gap constraint can be given by the user. To efficiently mine patterns, this paper proposes the ONP-Miner algorithm, which employs depth-first and backtracking strategies to calculate the support. Therefore, ONP-Miner can effectively avoid creating redundant nodes and parent-child relationships. Moreover, to effectively reduce the number of candidate patterns, ONP-Miner uses pattern join and pruning strategies to generate and further prune the candidate patterns, respectively. Experimental results show that ONP-Miner not only improves the mining efficiency, but also has better mining performance than the state-of-the-art algorithms. More importantly, ONP mining can find more interesting patterns in traffic volume data to predict future traffic.

cs.DB↗

Efficient Non-Autoregressive GAN Voice Conversion using VQWav2vec Features and Dynamic Convolution

It was shown recently that a combination of ASR and TTS models yield highly competitive performance on standard voice conversion tasks such as the Voice Conversion Challenge 2020 (VCC2020). To obtain good performance both models require pretraining on large amounts of data, thereby obtaining large models that are potentially inefficient in use. In this work we present a model that is significantly smaller and thereby faster in processing while obtaining equivalent performance. To achieve this the proposed model, Dynamic-GAN-VC (DYGAN-VC), uses a non-autoregressive structure and makes use of vector quantised embeddings obtained from a VQWav2vec model. Furthermore dynamic convolution is introduced to improve speech content modeling while requiring a small number of parameters. Objective and subjective evaluation was performed using the VCC2020 task, yielding MOS scores of up to 3.86, and character error rates as low as 4.3\%. This was achieved with approximately half the number of model parameters, and up to 8 times faster decoding speed.

eess.AS↗

On $\mathbb{F}_p$-roots of the Hilbert class polynomial modulo $p$

The Hilbert class polynomial $H_{\mathcal{O}}(x)\in \mathbb{Z}[x]$ attached to an order $\mathcal{O}$ in an imaginary quadratic field $K$ is the monic polynomial whose roots are precisely the distinct $j$-invariants of elliptic curves over $\mathbb{C}$ with complex multiplication by $\mathcal{O}$. Let $p$ be a prime inert in $K$ and strictly greater than $|\operatorname{disc}(\mathcal{O})|$. We show that the number of $\mathbb{F}_p$-roots of $H_\mathcal{O}(x)\!\! \pmod{p}$ is either zero or $|\operatorname{Pic}(\mathcal{O})[2]|$ by exhibiting a free and transitive action of $\operatorname{Pic}(\mathcal{O})[2]$ on the set of $\mathbb{F}_p$-roots of $H_\mathcal{O}(x)\!\! \pmod p$ whenever it is nonempty. We also provide a concrete criterion for the nonemptiness of the set of $\mathbb{F}_p$-roots. A similar result was first obtained by Xiao et al.~[Int. J. Number Theory, DOI: 10.1142/S1793042122500555] and generalized much further by Li et al.~[arXiv:2108.00168] (that covers the current result) with a different approach.

math.NT↗

Quadratic Chabauty for Atkin-Lehner Quotients of Modular Curves of Prime Level and Genus 4, 5, 6

We use the method of quadratic Chabauty on the quotients $X_0^+(N)$ of modular curves $X_0(N)$ by their Fricke involutions to provably compute all the rational points of these curves for prime levels $N$ of genus four, five, and six. We find that the only such curves with exceptional rational points are of levels $137$ and $311$. In particular there are no exceptional rational points on those curves of genus five and six. More precisely, we determine the rational points on the curves $X_0^+(N)$ for $N=137,173,199,251,311,157,181,227,263,163,197,211,223,269,271,359$.

math.NT↗

Towards Low-Resource StarGAN Voice Conversion using Weight Adaptive Instance Normalization

Many-to-many voice conversion with non-parallel training data has seen significant progress in recent years. StarGAN-based models have been interests of voice conversion. However, most of the StarGAN-based methods only focused on voice conversion experiments for the situations where the number of speakers was small, and the amount of training data was large. In this work, we aim at improving the data efficiency of the model and achieving a many-to-many non-parallel StarGAN-based voice conversion for a relatively large number of speakers with limited training samples. In order to improve data efficiency, the proposed model uses a speaker encoder for extracting speaker embeddings and conducts adaptive instance normalization (AdaIN) on convolutional weights. Experiments are conducted with 109 speakers under two low-resource situations, where the number of training samples is 20 and 5 per speaker. An objective evaluation shows the proposed model is better than the baseline methods. Furthermore, a subjective evaluation shows that, for both naturalness and similarity, the proposed model outperforms the baseline method.

cs.SD↗

Full-color nanorouter for high-resolution imaging

Pixel scaling effects have been a major issue for the development of high-resolution color image sensors due to the reduced photoelectric signal and the color crosstalk. Various structural color techniques have been proposed and demonstrated the large freedom in color manipulation by the structure design. However, the optical efficiency and the color distortion limit the practical applications due to its intrinsic filtering mechanism. Instead, on-chip full-color routing is quite charming for improving the signal-to-noise ratio. In this paper, a single-layer quick response code like nanorouter is proposed for the full-color light routing in a pixel level of image sensors. It shows much higher routing efficiency than various planar lens schemes for signal wavelength focusing. Moreover, over 60% signal enhancement with robust polarization insensitivity is obtained in all three primary color bands with a same nanorouter by a multi-objective optimization method. Negligible color distortion is observed from the reconstructed color image. Such a simple nanorouter structure is promising for the development of image sensor, photovoltaics and display.

physics.optics↗

T-vectors: Weakly Supervised Speaker Identification Using Hierarchical Transformer Model

Identifying multiple speakers without knowing where a speaker's voice is in a recording is a challenging task. This paper proposes a hierarchical network with transformer encoders and memory mechanism to address this problem. The proposed model contains a frame-level encoder and segment-level encoder, both of them make use of the transformer encoder block. The multi-head attention mechanism in the transformer structure could better capture different speaker properties when the input utterance contains multiple speakers. The memory mechanism used in the frame-level encoders can build a recurrent connection that better capture long-term speaker features. The experiments are conducted on artificial datasets based on the Switchboard Cellular part1 (SWBC) and Voxceleb1 datasets. In different data construction scenarios (Concat and Overlap), the proposed model shows better performance comparaing with four strong baselines, reaching 13.3% and 10.5% relative improvement compared with H-vectors and S-vectors. The use of memory mechanism could reach 10.6% and 7.7% relative improvement compared with not using memory mechanism.

cs.SD↗

Unsupervised Acoustic Unit Representation Learning for Voice Conversion using WaveNet Auto-encoders

Unsupervised representation learning of speech has been of keen interest in recent years, which is for example evident in the wide interest of the ZeroSpeech challenges. This work presents a new method for learning frame level representations based on WaveNet auto-encoders. Of particular interest in the ZeroSpeech Challenge 2019 were models with discrete latent variable such as the Vector Quantized Variational Auto-Encoder (VQVAE). However these models generate speech with relatively poor quality. In this work we aim to address this with two approaches: first WaveNet is used as the decoder and to generate waveform data directly from the latent representation; second, the low complexity of latent representations is improved with two alternative disentanglement learning methods, namely instance normalization and sliced vector quantization. The method was developed and tested in the context of the recent ZeroSpeech challenge 2020. The system output submitted to the challenge obtained the top position for naturalness (Mean Opinion Score 4.06), top position for intelligibility (Character Error Rate 0.15), and third position for the quality of the representation (ABX test score 12.5). These and further analysis in this paper illustrates that quality of the converted speech and the acoustic units representation can be well balanced.

eess.AS↗