Search

One-shot voice conversion(VC) aims to change the timbre of any source speech to match that of the target speaker with only one speech sample. Existing style transfer-based VC methods relied on speech representation disentanglement and suffered from accurately and independently encoding each speech component and recomposing back to converted speech effectively. To tackle this, we proposed Pureformer-VC, which utilizes Conformer blocks to build a disentangled encoder, and Zipformer blocks to build a style transfer decoder as the generator. In the decoder, we used effective styleformer blocks to integrate speaker characteristics effectively into the generated speech. The models used the generative VAE loss for encoding components and triplet loss for unsupervised discriminative training. We applied the styleformer method to Zipformer's shared weights for style transfer. The experimental results show that the proposed model achieves comparable subjective scores and exhibits improvements in objective metrics compared to existing methods in a one-shot voice conversion scenario., Comment: submmited to ICASSP 2025

Deep speech classification tasks, mainly including keyword spotting and speaker verification, play a crucial role in speech-based human-computer interaction. Recently, the security of these technologies has been demonstrated to be vulnerable to backdoor attacks. Specifically speaking, speech samples are attacked by noisy disruption and component modification in present triggers. We suggest that speech backdoor attacks can strategically focus on emotion, a higher-level subjective perceptual attribute inherent in speech. Furthermore, we proposed that emotional voice conversion technology can serve as the speech backdoor attack trigger, and the method is called EmoAttack. Based on this, we conducted attack experiments on two speech classification tasks, showcasing that EmoAttack method owns impactful trigger effectiveness and its remarkable attack success rate and accuracy variance. Additionally, the ablation experiments found that speech with intensive emotion is more suitable to be targeted for attacks., Comment: Submitted to ICASSP 2025

Speech recognition is an essential start ring of human-computer interaction, and recently, deep learning models have achieved excellent success in this task. However, when the model training and private data provider are always separated, some security threats that make deep neural networks (DNNs) abnormal deserve to be researched. In recent years, the typical backdoor attacks have been researched in speech recognition systems. The existing backdoor methods are based on data poisoning. The attacker adds some incorporated changes to benign speech spectrograms or changes the speech components, such as pitch and timbre. As a result, the poisoned data can be detected by human hearing or automatic deep algorithms. To improve the stealthiness of data poisoning, we propose a non-neural and fast algorithm called Random Spectrogram Rhythm Transformation (RSRT) in this paper. The algorithm combines four steps to generate stealthy poisoned utterances. From the perspective of rhythm component transformation, our proposed trigger stretches or squeezes the mel spectrograms and recovers them back to signals. The operation keeps timbre and content unchanged for good stealthiness. Our experiments are conducted on two kinds of speech recognition tasks, including testing the stealthiness of poisoned samples by speaker verification and automatic speech recognition. The results show that our method has excellent effectiveness and stealthiness. The rhythm trigger needs a low poisoning rate and gets a very high attack success rate.

Regular expressions (regexes) are widely used in different fields of computer science, such as programming languages, string processing, and databases. However, existing tools for synthesizing or repairing regexes always assume that the input examples are faultless. In real industrial scenarios, this assumption does not entirely hold. Thus, this paper presents a simple but effective templated-based approach to generate regular expressions over noisy examples. Specifically, we present a data model (i.e., MetaParam) to extract features of strings for clustering all examples. Then, we propose a practical dynamic thresholding scheme to filter out anomalous examples via detecting knee points on CDF graphs. Finally, we design a template-based algorithm to translate a finite of positve examples to regular expression, which is efficient, interpretable, and extensible. We performed an experimental evaluation on four different extraction tasks applied to real-world datasets and obtained promising results in terms of F-measure. Moreover, gMeta achieves excellent results in real industrial scenarios.

We investigate the properties of long tidal tails using the largest to date sample of 461 merging galaxies with $\log(M_\ast/\rm M_\odot)\geq9.5$ within $0.2 \leq z \leq 1$ from the COSMOS survey in combination with {\it Hubble Space Telescope} imaging data. Long tidal tails can be briefly divided into three shape types: straight (41\,per\,cent), curved (47\,per\,cent) and plume (12\,per\,cent). Their host galaxies are mostly at late stages of merging, although 31\,per\,cent are galaxy pairs with projected separations $d>20$\,kpc. The high formation rate of straight tidal tails needs to be understood as the projection of curved tidal tails accounts for only a small fraction of the straight tails. We identify 165 tidal dwarf galaxies (TDGs), yielding a TDG production rate of 0.36 per merger. Combined with a galaxy merger fraction and a TDG survival rate from the literature, we estimate that $\sim$\,5\,per\,cent of local dwarf galaxies are of tidal origin, suggesting the tidal formation is not an important formation channel for the dwarf galaxies. More than half of TDGs are located at the tip of their host tails. These TDGs have stellar masses in the range of $7.5\leq\log (M_\ast/\rm M_\odot)\leq9.5$ and appear compact with half-light radii following the $M_\ast$ - $R_{\rm e}$ relation of low-mass elliptical galaxies. However, their surface brightness profiles are generally flatter than those of local disc galaxies. Only 10 out of 165 TDGs have effective radii larger than 1.5 kpc and would qualify as unusually bright ultra-diffuse galaxies., Comment: 11 pages, 13 figures, accepted for publication in MNRAS

As architecture, systems, and data management communities pay greater attention to innovative big data systems and architectures, the pressure of benchmarking and evaluating these systems rises. Considering the broad use of big data systems, big data benchmarks must include diversity of data and workloads. Most of the state-of-the-art big data benchmarking efforts target evaluating specific types of applications or system software stacks, and hence they are not qualified for serving the purposes mentioned above. This paper presents our joint research efforts on this issue with several industrial partners. Our big data benchmark suite BigDataBench not only covers broad application scenarios, but also includes diverse and representative data sets. BigDataBench is publicly available from http://prof.ict.ac.cn/BigDataBench . Also, we comprehensively characterize 19 big data workloads included in BigDataBench with varying data inputs. On a typical state-of-practice processor, Intel Xeon E5645, we have the following observations: First, in comparison with the traditional benchmarks: including PARSEC, HPCC, and SPECCPU, big data applications have very low operation intensity; Second, the volume of data input has non-negligible impact on micro-architecture characteristics, which may impose challenges for simulation-based big data architecture research; Last but not least, corroborating the observations in CloudSuite and DCBench (which use smaller data inputs), we find that the numbers of L1 instruction cache misses per 1000 instructions of the big data applications are higher than in the traditional benchmarks; also, we find that L3 caches are effective for the big data applications, corroborating the observation in DCBench., Comment: 12 pages, 6 figures, The 20th IEEE International Symposium On High Performance Computer Architecture (HPCA-2014), February 15-19, 2014, Orlando, Florida, USA

This paper presents our joint research efforts on big data benchmarking with several industrial partners. Considering the complexity, diversity, workload churns, and rapid evolution of big data systems, we take an incremental approach in big data benchmarking. For the first step, we pay attention to search engines, which are the most important domain in Internet services in terms of the number of page views and daily visitors. However, search engine service providers treat data, applications, and web access logs as business confidentiality, which prevents us from building benchmarks. To overcome those difficulties, with several industry partners, we widely investigated the open source solutions in search engines, and obtained the permission of using anonymous Web access logs. Moreover, with two years' great efforts, we created a sematic search engine named ProfSearch (available from http://prof.ict.ac.cn). These efforts pave the path for our big data benchmark suite from search engines---BigDataBench, which is released on the web page (http://prof.ict.ac.cn/BigDataBench). We report our detailed analysis of search engine workloads, and present our benchmarking methodology. An innovative data generation methodology and tool are proposed to generate scalable volumes of big data from a small seed of real data, preserving semantics and locality of data. Also, we preliminarily report two case studies using BigDataBench for both system and architecture researches., Comment: 7 pages, 5 figures, The Third Workshop on Architectures and Systems for Big Data(ASBD 2013) in conjunction with The 40th International Symposium on Computer Architecture, May 2013