arXiv subjects

Hongyu Xiang

Publications and source records attributed to Hongyu Xiang.

16 recordsLinked to original sources

arXiv2025

Isomorphism between Hopf algebras for multiple zeta values

The classical quasi-shuffle algebra for multiple zeta values have a well-known Hopf algebra structure. Recently, the shuffle algebra for multiple zeta values are also equipped with a Hopf algebra structure. This paper shows that these two Hopf algebras are isomorphic utilizing quasi-symmetric functions. This Hopf algebra isomorphism is compared with with the well-known isomorphism between the shuffle Hopf algebra and quasi-shuffle Hopf algebra of Hoffman, Newman and Radford.

math.NT↗

arXiv2025

LongCat-Flash-Omni Technical Report

We introduce LongCat-Flash-Omni, a state-of-the-art open-source omni-modal model with 560 billion parameters, excelling at real-time audio-visual interaction. By adopting a curriculum-inspired progressive training strategy that transitions from simpler to increasingly complex modality sequence modeling tasks, LongCat-Flash-Omni attains comprehensive multimodal capabilities while maintaining strong unimodal capability. Building upon LongCat-Flash, which adopts a high-performance Shortcut-connected Mixture-of-Experts (MoE) architecture with zero-computation experts, LongCat-Flash-Omni integrates efficient multimodal perception and speech reconstruction modules. Despite its immense size of 560B parameters (with 27B activated), LongCat-Flash-Omni achieves low-latency real-time audio-visual interaction. For training infrastructure, we developed a modality-decoupled parallelism scheme specifically designed to manage the data and model heterogeneity inherent in large-scale multimodal training. This innovative approach demonstrates exceptional efficiency by sustaining over 90% of the throughput achieved by text-only training. Extensive evaluations show that LongCat-Flash-Omni achieves state-of-the-art performance on omni-modal benchmarks among open-source models. Furthermore, it delivers highly competitive results across a wide range of modality-specific tasks, including text, image, and video understanding, as well as audio understanding and generation. We provide a comprehensive overview of the model architecture design, training procedures, and data strategies, and open-source the model to foster future research and development in the community.

cs.MM↗

arXiv2025

LongCat-Audio-Codec: An Audio Tokenizer and Detokenizer Solution Designed for Speech Large Language Models

This paper presents LongCat-Audio-Codec, an audio tokenizer and detokenizer solution designed for industrial grade end-to-end speech large language models. By leveraging a decoupled model architecture and a multistage training strategy, LongCat-Audio-Codec exhibits robust semantic modeling capabilities, flexible acoustic feature extraction capabilities, and low-latency streaming synthesis capabilities. It encodes speech at an ultra-low frame rate of 16.67 Hz, with a minimum bitrate of 0.43 kbps and a maximum bitrate of 0.87 kbps. Evaluation results demonstrate that LongCat-Audio-Codec achieves strong speech intelligibility and is capable of synthesizing highquality speech at low bitrate, thus effectively balancing coding efficiency and decoding quality. The inference code and model checkpoints of LongCat-Audio-Codec are available at: https://github.com/meituan-longcat/LongCat-Audio-Codec.

eess.AS↗

arXiv2024

A Riemann-type duality of shuffle Hopf algebras related to multiple zeta values

This paper offers a Hopf algebraic interpretation of a functional equation of multiple zeta functions, motivated by the classical symmetry of the Riemann zeta function. Starting from the extended shuffle algebra that encodes multiple zeta values (MZVs) at integer arguments, we show that its subalgebra corresponding to nonpositive arguments carries a natural differential Hopf algebra structure. This Hopf algebra is in graded linear duality with the shuffle Hopf algebra associated to MZVs at positive arguments. The resulting duality, realized through an explicit isomorphism, provides an algebraic analog of the functional equation relating $\zeta(s)$ with $\zeta(1-s)$ of the Riemann zeta function and unifies the positive and nonpositive sectors of multiple zeta functions within a common Hopf algebraic framework.

math.RA↗

arXiv2024

Extended Shuffle Product for Multiple Zeta Values

The shuffle algebra on positive integers encodes the usual multiple zeta values (MZVs) (with positive arguments) thanks to the representations of MZVs by iterated Chen integrals of Kontsevich. Together with the quasi-shuffle (stuffle) algebra, it provides the algebraic framework to study relations among MZVs. This paper enlarges the shuffle algebra uniquely to what we call the extended shuffle algebra that encodes convergent multiple zeta series with arbitrary integer arguments, not just the positive ones in the usual case. To achieved this goal, we first replace the Rota-Baxter operator of weight zero (the integral operator) that characterizes the shuffle product by the differential operator which extends the shuffle product to the larger space. We then show that the subspace corresponding to the convergent MZVs with integer arguments becomes a subalgebra under this extended shuffle product. Furthermore, by lifting the extended shuffle algebra to the locality algebra of Chen symbols, we prove that taking summations of fractions from Chen symbols defines an algebra homomorphism from the above subalgebra to the subalgebra of real numbers spanned by convergent multiple zeta series.

math.NT↗

arXiv2023

Hopf algebras for the shuffle algebra and fractions from multiple zeta values

The algebra of multiple zeta values (MZVs) is encoded as a stuffle (quasi-shuffle) algebra and a shuffle algebra. The MZV stuffle algebra has a natural Hopf algebra structure. This paper equips a Hopf algebra structure to the MZV shuffle algebra. The needed coproduct is defined by a recursion through a family of weight-increasing linear operators. To verify the Hopf algebra axioms, we make use of a family of fractions, called Chen fractions, that have been used to study MZVs and also serve as the function model for the MZV shuffle algebra. Applying natural derivations on functions and working in the context of locality, a locality Hopf algebra structure is established on the linear span of Chen fractions. This locality Hopf algebra is then shown to descend to a Hopf algebra on the MZV shuffle algebra, whose coproduct satisfies the same recursion as the first-defined coproduct. Thus the two coproducts coincide, establishing the needed Hopf algebra axioms on the MZV shuffle algebra.

math.NT↗

arXiv2023

CPPF: A contextual and post-processing-free model for automatic speech recognition

ASR systems have become increasingly widespread in recent years. However, their textual outputs often require post-processing tasks before they can be practically utilized. To address this issue, we draw inspiration from the multifaceted capabilities of LLMs and Whisper, and focus on integrating multiple ASR text processing tasks related to speech recognition into the ASR model. This integration not only shortens the multi-stage pipeline, but also prevents the propagation of cascading errors, resulting in direct generation of post-processed text. In this study, we focus on ASR-related processing tasks, including Contextual ASR and multiple ASR post processing tasks. To achieve this objective, we introduce the CPPF model, which offers a versatile and highly effective alternative to ASR processing. CPPF seamlessly integrates these tasks without any significant loss in recognition performance.

cs.CL↗

arXiv2022

Peak-First CTC: Reducing the Peak Latency of CTC Models by Applying Peak-First Regularization

The CTC model has been widely applied to many application scenarios because of its simple structure, excellent performance, and fast inference speed. There are many peaks in the probability distribution predicted by the CTC models, and each peak represents a non-blank token. The recognition latency of CTC models can be reduced by encouraging the model to predict peaks earlier. Existing methods to reduce latency require modifying the transition relationship between tokens in the forward-backward algorithm, and the gradient calculation. Some of these methods even depend on the forced alignment results provided by other pretrained models. The above methods are complex to implement. To reduce the peak latency, we propose a simple and novel method named peak-first regularization, which utilizes a frame-wise knowledge distillation function to force the probability distribution of the CTC model to shift left along the time axis instead of directly modifying the calculation process of CTC loss and gradients. All the experiments are conducted on a Chinese Mandarin dataset AISHELL-1. We have verified the effectiveness of the proposed regularization on both streaming and non-streaming CTC models respectively. The results show that the proposed method can reduce the average peak latency by about 100 to 200 milliseconds with almost no degradation of recognition accuracy.

eess.AS↗

arXiv2022

CUSIDE: Chunking, Simulating Future Context and Decoding for Streaming ASR

History and future contextual information are known to be important for accurate acoustic modeling. However, acquiring future context brings latency for streaming ASR. In this paper, we propose a new framework - Chunking, Simulating Future Context and Decoding (CUSIDE) for streaming speech recognition. A new simulation module is introduced to recursively simulate the future contextual frames, without waiting for future context. The simulation module is jointly trained with the ASR model using a self-supervised loss; the ASR model is optimized with the usual ASR loss, e.g., CTC-CRF as used in our experiments. Experiments show that, compared to using real future frames as right context, using simulated future context can drastically reduce latency while maintaining recognition accuracy. With CUSIDE, we obtain new state-of-the-art streaming ASR results on the AISHELL-1 dataset.

eess.AS↗

arXiv2020

CAT: A CTC-CRF based ASR Toolkit Bridging the Hybrid and the End-to-end Approaches towards Data Efficiency and Low Latency

In this paper, we present a new open source toolkit for speech recognition, named CAT (CTC-CRF based ASR Toolkit). CAT inherits the data-efficiency of the hybrid approach and the simplicity of the E2E approach, providing a full-fledged implementation of CTC-CRFs and complete training and testing scripts for a number of English and Chinese benchmarks. Experiments show CAT obtains state-of-the-art results, which are comparable to the fine-tuned hybrid models in Kaldi but with a much simpler training pipeline. Compared to existing non-modularized E2E models, CAT performs better on limited-scale datasets, demonstrating its data efficiency. Furthermore, we propose a new method called contextualized soft forgetting, which enables CAT to do streaming ASR without accuracy degradation. We hope CAT, especially the CTC-CRF based framework and software, will be of broad interest to the community, and can be further explored and improved.

eess.AS↗

arXiv2020

Mode Selection and Resource Allocation in Sliced Fog Radio Access Networks: A Reinforcement Learning Approach

The mode selection and resource allocation in fog radio access networks (F-RANs) have been advocated as key techniques to improve spectral and energy efficiency. In this paper, we investigate the joint optimization of mode selection and resource allocation in uplink F-RANs, where both of the traditional user equipments (UEs) and fog UEs are served by constructed network slice instances. The concerned optimization is formulated as a mixed-integer programming problem, and both the orthogonal and multiplexed subchannel allocation strategies are proposed to guarantee the slice isolation. Motivated by the development of machine learning, two reinforcement learning based algorithms are developed to solve the original high complexity problem under traditional and fog UEs' specific performance requirements. The basic idea of the proposals is to generate a good mode selection policy according to the immediate reward fed back by an environment. Simulation results validate the benefits of our proposed algorithms and show that a tradeoff between system power consumption and queue delay can be achieved.

cs.NI↗

arXiv2019

CAT: CRF-based ASR Toolkit

In this paper, we present a new open source toolkit for automatic speech recognition (ASR), named CAT (CRF-based ASR Toolkit). A key feature of CAT is discriminative training in the framework of conditional random field (CRF), particularly with connectionist temporal classification (CTC) inspired state topology. CAT contains a full-fledged implementation of CTC-CRF and provides a complete workflow for CRF-based end-to-end speech recognition. Evaluation results on Chinese and English benchmarks such as Switchboard and Aishell show that CAT obtains the state-of-the-art results among existing end-to-end models with less parameters, and is competitive compared with the hybrid DNN-HMM models. Towards flexibility, we show that i-vector based speaker-adapted recognition and latency control mechanism can be explored easily and effectively in CAT. We hope CAT, especially the CRF-based framework and software, will be of broad interest to the community, and can be further explored and improved.

cs.LG↗

arXiv2017

Network Slicing in Fog Radio Access Networks: Issues and Challenges

Network slicing has been advocated by both academia and industry as a cost-efficient way to enable operators to provide networks on an as-a-service basis and meet the wide range of use cases that the fifth generation wireless network will serve. The existing works on network slicing are mainly targeted at the partition of the core network, and the prospect of network slicing in radio access networks should be jointly exploited. To solve this challenge, an enhanced network slicing in fog radio access networks (F-RANs), termed as access slicing, is proposed. This article comprehensively presents a novel architecture and related key techniques for access slicing in F-RANs. The proposed hierarchical architecture of access slicing consists of centralized orchestration layer and slice instance layer, which makes the access slicing adaptively implement in an convenient way. Meanwhile, key techniques and their corresponding solutions, including the radio and cache resource management, as well as the social-aware slicing, are presented. Open issues in terms of standardization developments and field trials are identified.

cs.IT↗

arXiv2017

Resource Allocation in Cloud Radio Access Networks with Device-to-Device Communications

To alleviate the burdens on the fronthaul and reduce the transmit latency, the device-to-device (D2D) communication is presented in cloud radio access networks (C-RANs). Considering dynamic traffic arrivals and time-varying channel conditions, the resource allocation in C-RANs with D2D is formulated into a stochastic optimization problem, which is aimed at maximizing the overall throughput subject to network stability, interference, and fronthaul capacity constraints. Leveraging on the Lyapunov optimization technique, the stochastic optimization problem is transformed into a delay-aware optimization problem, which is a mixed-integer nonlinear programming problem and can be decomposed into three subproblems: mode selection, uplink beamforming design, and power control. An optimization solution that consists of a modified branch and bound method as well as a weighted minimum mean square error approach has been developed to obtain the close-to-optimal solution. Simulation results validate that the D2D can improve throughput, decrease latency, and alleviate the burdens of the constrained fronthaul in C-RANs. Furthermore, an average throughput-delay tradeoff can be achieved by the proposed solution.

cs.IT↗

arXiv2016

Inter-tier Interference Suppression in Heterogeneous Cloud Radio Access Networks

Incorporating cloud computing into heterogeneous networks, the heterogeneous cloud radio access network (H-CRAN) has been proposed as a promising paradigm to enhance both spectral and energy efficiencies. Developing interference suppression strategies is critical for suppressing the inter-tier interference between remote radio heads (RRHs) and a macro base station (MBS) in H-CRANs. In this paper, inter-tier interference suppression techniques are considered in the contexts of collaborative processing and cooperative radio resource allocation (CRRA). In particular, interference collaboration (IC) and beamforming (BF) are proposed to suppress the inter-tier interference, and their corresponding performance is evaluated. Closed-form expressions for the overall outage probabilities, system capacities, and average bit error rates under these two schemes are derived. Furthermore, IC and BF based CRRA optimization models are presented to maximize the RRH-accessed users' sum rates via power allocation, which is solved with convex optimization. Simulation results demonstrate that the derived expressions for these performance metrics for IC and BF are accurate; and the relative performance between IC and BF schemes depends on system parameters, such as the number of antennas at the MBS, the number of RRHs, and the target signal-to-interference-plus-noise ratio threshold. Furthermore, it is seen that the sum rates of IC and BF schemes increase almost linearly with the transmit power threshold under the proposed CRRA optimization solution.

cs.IT↗

arXiv2016

Energy-Efficient Resource Allocation Optimization for Multimedia Heterogeneous Cloud Radio Access Networks

The heterogeneous cloud radio access network (H-CRAN) is a promising paradigm which incorporates the cloud computing into heterogeneous networks (HetNets), thereby taking full advantage of cloud radio access networks (C-RANs) and HetNets. Characterizing the cooperative beamforming with fronthaul capacity and queue stability constraints is critical for multimedia applications to improving energy efficiency (EE) in H-CRANs. An energy-efficient optimization objective function with individual fronthaul capacity and inter-tier interference constraints is presented in this paper for queue-aware multimedia H-CRANs. To solve this non-convex objective function, a stochastic optimization problem is reformulated by introducing the general Lyapunov optimization framework. Under the Lyapunov framework, this optimization problem is equivalent to an optimal network-wide cooperative beamformer design algorithm with instantaneous power, average power and inter-tier interference constraints, which can be regarded as the weighted sum EE maximization problem and solved by a generalized weighted minimum mean square error approach. The mathematical analysis and simulation results demonstrate that a tradeoff between EE and queuing delay can be achieved, and this tradeoff strictly depends on the fronthaul constraint.

cs.NI↗