Accents, as variations from standard pronunciation, pose significant challenges for speech recognition systems. Although joint automatic speech recognition (ASR) and accent recognition (AR) training has been proven effective in handling multi-accent scenarios, current multi-task ASR-AR approaches overlook the granularity differences between tasks. Fine-grained units capture pronunciation-related accent characteristics, while coarse-grained units are better for learning linguistic information. Moreover, an explicit interaction of two tasks can also provide complementary information and improve the performance of each other, but it is rarely used by existing approaches. In this paper, we propose a novel Decoupling and Interacting Multi-task Network (DIMNet) for joint speech and accent recognition, which is comprised of a connectionist temporal classification (CTC) branch, an AR branch, an ASR branch, and a bottom feature encoder. Specifically, AR and ASR are first decoupled by separated branches and two-granular modeling units to learn task-specific representations. The AR branch is from our previously proposed linguistic-acoustic bimodal AR model and the ASR branch is an encoder-decoder based Conformer model. Then, for the task interaction, the CTC branch provides aligned text for the AR task, while accent embeddings extracted from our AR model are incorporated into the ASR branch's encoder and decoder. Finally, during ASR inference, a cross-granular rescoring method is introduced to fuse the complementary information from the CTC and attention decoder after the decoupling. Our experiments on English and Chinese datasets demonstrate the effectiveness of the proposed model, which achieves 21.45%/28.53% AR accuracy relative improvement and 32.33%/14.55% ASR error rate relative reduction over a published standard baseline, respectively.
相關內容
The joint progress of artificial neural networks (ANNs) and domain specific hardware accelerators such as GPUs and TPUs took over many domains of machine learning research. This development is accompanied by a rapid growth of the required computational demands for larger models and more data. Concurrently, emerging properties of foundation models such as in-context learning drive new opportunities for machine learning applications. However, the computational cost of such applications is a limiting factor of the technology in data centers, and more importantly in mobile devices and edge systems. To mediate the energy footprint and non-trivial latency of contemporary systems, neuromorphic computing systems deeply integrate computational principles of neurobiological systems by leveraging low-power analog and digital technologies. SpiNNaker2 is a digital neuromorphic chip developed for scalable machine learning. The event-based and asynchronous design of SpiNNaker2 allows the composition of large-scale systems involving thousands of chips. This work features the operating principles of SpiNNaker2 systems, outlining the prototype of novel machine learning applications. These applications range from ANNs over bio-inspired spiking neural networks to generalized event-based neural networks. With the successful development and deployment of SpiNNaker2, we aim to facilitate the advancement of event-based and asynchronous algorithms for future generations of machine learning systems.
ChatGPT has piqued the interest of many fields, particularly in the academic community. GPT-4, the latest version, starts supporting multimodal input and output. This study examines social media posts to analyze how the Chinese public perceives the potential of ChatGPT for educational and general purposes. The study also serves as the first effort to investigate the changes in public opinion since the release of GPT-4. According to the analysis results, prior to GPT-4, although some social media users believed that AI advancements would benefit education and society, some believed that advanced AI, such as ChatGPT, would make humans feel inferior and lead to problems such as cheating and a decline in moral principles, while the majority remain neutral. Interestingly, public attitudes have tended to shift in a positive direction since the release of GPT-4. We present a thorough analysis of the trending shift and a roadmap to ensure the ethical application of ChatGPT-like models in education and beyond.
Sophisticated cyber attacks present significant challenges for organizations in detecting and preventing such threats. To address this critical need for advanced defense mechanisms, we propose an Ensemble Defense System (EDS). An EDS is a cybersecurity framework aggregating multiple security tools designed to monitor and alert an organization during cyber attacks. The proposed EDS leverages a comprehensive range of Intrusion Detection System (IDS) capabilities by introducing a hybrid of signature-based IDS and anomaly-based IDS tools. It also incorporates Elasticsearch, an open-source Security Information and Event Management (SIEM) tool, to facilitate data analysis and interactive visualization of alerts generated from IDSs. The effectiveness of the EDS is evaluated through a payload from a bash script that executes various attacks, including port scanning, privilege escalation, and Denial-of-Service (DoS). The evaluation demonstrates the EDS's ability to detect diverse cyber attacks.
We introduce two data-driven procedures for optimal estimation and inference in nonparametric models using instrumental variables. The first is a data-driven choice of sieve dimension for a popular class of sieve two-stage least squares estimators. When implemented with this choice, estimators of both the structural function $h_0$ and its derivatives (such as elasticities) converge at the fastest possible (i.e., minimax) rates in sup-norm. The second is for constructing uniform confidence bands (UCBs) for $h_0$ and its derivatives. Our UCBs guarantee coverage over a generic class of data-generating processes and contract at the minimax rate, possibly up to a logarithmic factor. As such, our UCBs are asymptotically more efficient than UCBs based on the usual approach of undersmoothing. As an application, we estimate the elasticity of the intensive margin of firm exports in a monopolistic competition model of international trade. Simulations illustrate the good performance of our procedures in empirically calibrated designs. Our results provide evidence against common parameterizations of the distribution of unobserved firm heterogeneity.
We present Modular Polynomial (MP) Codes for Secure Distributed Matrix Multiplication (SDMM). The construction is based on the observation that one can decode certain proper subsets of the coefficients of a polynomial with fewer evaluations than is necessary to interpolate the entire polynomial. We also present Generalized Gap Additive Secure Polynomial (GGASP) codes. Both MP and GGASP codes are shown experimentally to perform favorably in terms of recovery threshold when compared to other comparable polynomials codes for SDMM which use the grid partition. Both MP and GGASP codes achieve the recovery threshold of Entangled Polynomial Codes for robustness against stragglers, but MP codes can decode below this recovery threshold depending on the set of worker nodes which fails. The decoding complexity of MP codes is shown to be lower than other approaches in the literature, due to the user not being tasked with interpolating an entire polynomial.
Large language models (LLMs) have made significant advancements in code-related tasks, yet many LLMs treat code as simple sequences, neglecting its structured nature. We introduce AST-T5, a novel pretraining paradigm that leverages the Abstract Syntax Tree (AST) for enhanced code generation, transpilation, and understanding. Using dynamic programming, our AST-Aware Segmentation retains code structure, while our AST-Aware Span Corruption objective equips the model to reconstruct various code structures. Unlike other models, AST-T5 avoids intricate program analyses or architectural changes, so it integrates seamlessly with any encoder-decoder Transformer. Evaluations show that AST-T5 consistently outperforms similar-sized LMs across various code-related tasks. Structure-awareness makes AST-T5 particularly powerful in code-to-code tasks, surpassing CodeT5 by 2 points in exact match score for the Bugs2Fix task and by 3 points in exact match score for Java-C# Transpilation in CodeXGLUE. Our code and model are publicly available at //github.com/gonglinyuan/ast_t5.
Optical lithography is the main enabler to semiconductor manufacturing. It requires extensive processing to perform the Resolution Enhancement Techniques (RETs) required to transfer the design data to a working Integrated Circuits (ICs). The processing power and computational runtime for RETs tasks is ever increasing due to the continuous reduction of the feature size and the expansion of the chip area. State-of-the-art research sought Machine Learning (ML) technologies to reduce runtime and computational power, however they are still not used in production yet. In this study, we analyze the reasons holding back ML computational lithography from being production ready and present a novel highly scalable end-to-end flow that enables production ready ML-RET correction.
Existing knowledge graph (KG) embedding models have primarily focused on static KGs. However, real-world KGs do not remain static, but rather evolve and grow in tandem with the development of KG applications. Consequently, new facts and previously unseen entities and relations continually emerge, necessitating an embedding model that can quickly learn and transfer new knowledge through growth. Motivated by this, we delve into an expanding field of KG embedding in this paper, i.e., lifelong KG embedding. We consider knowledge transfer and retention of the learning on growing snapshots of a KG without having to learn embeddings from scratch. The proposed model includes a masked KG autoencoder for embedding learning and update, with an embedding transfer strategy to inject the learned knowledge into the new entity and relation embeddings, and an embedding regularization method to avoid catastrophic forgetting. To investigate the impacts of different aspects of KG growth, we construct four datasets to evaluate the performance of lifelong KG embedding. Experimental results show that the proposed model outperforms the state-of-the-art inductive and lifelong embedding baselines.
Automatic KB completion for commonsense knowledge graphs (e.g., ATOMIC and ConceptNet) poses unique challenges compared to the much studied conventional knowledge bases (e.g., Freebase). Commonsense knowledge graphs use free-form text to represent nodes, resulting in orders of magnitude more nodes compared to conventional KBs (18x more nodes in ATOMIC compared to Freebase (FB15K-237)). Importantly, this implies significantly sparser graph structures - a major challenge for existing KB completion methods that assume densely connected graphs over a relatively smaller set of nodes. In this paper, we present novel KB completion models that can address these challenges by exploiting the structural and semantic context of nodes. Specifically, we investigate two key ideas: (1) learning from local graph structure, using graph convolutional networks and automatic graph densification and (2) transfer learning from pre-trained language models to knowledge graphs for enhanced contextual representation of knowledge. We describe our method to incorporate information from both these sources in a joint model and provide the first empirical results for KB completion on ATOMIC and evaluation with ranking metrics on ConceptNet. Our results demonstrate the effectiveness of language model representations in boosting link prediction performance and the advantages of learning from local graph structure (+1.5 points in MRR for ConceptNet) when training on subgraphs for computational efficiency. Further analysis on model predictions shines light on the types of commonsense knowledge that language models capture well.
High spectral dimensionality and the shortage of annotations make hyperspectral image (HSI) classification a challenging problem. Recent studies suggest that convolutional neural networks can learn discriminative spatial features, which play a paramount role in HSI interpretation. However, most of these methods ignore the distinctive spectral-spatial characteristic of hyperspectral data. In addition, a large amount of unlabeled data remains an unexploited gold mine for efficient data use. Therefore, we proposed an integration of generative adversarial networks (GANs) and probabilistic graphical models for HSI classification. Specifically, we used a spectral-spatial generator and a discriminator to identify land cover categories of hyperspectral cubes. Moreover, to take advantage of a large amount of unlabeled data, we adopted a conditional random field to refine the preliminary classification results generated by GANs. Experimental results obtained using two commonly studied datasets demonstrate that the proposed framework achieved encouraging classification accuracy using a small number of data for training.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191