Cong Zhou

Yong Dai, Linyang Li, Cong Zhou et al.

Whole word masking (WWM), which masks all subwords corresponding to a word at once, makes a better English BERT model. For the Chinese language, however, there is no subword because each token is an atomic character. The meaning of a word in Chinese is different in that a word is a compositional unit consisting of multiple characters. Such difference motivates us to investigate whether WWM leads to better context understanding ability for Chinese BERT. To achieve this, we introduce two probing tasks related to grammatical error correction and ask pretrained models to revise or insert tokens in a masked language modeling manner. We construct a dataset including labels for 19,075 tokens in 10,448 sentences. We train three Chinese BERT models with standard character-level masking (CLM), WWM, and a combination of CLM and WWM, respectively. Our major findings are as follows: First, when one character needs to be inserted or replaced, the model trained with CLM performs the best. Second, when more than one character needs to be handled, WWM is the key to better performance. Finally, when being fine-tuned on sentence-level downstream tasks, models trained with different masking strategies perform comparably.

Yong Dai, Duyu Tang, Liangxin Liu et al.

People perceive the world with multiple senses (e.g., through hearing sounds, reading words and seeing objects). However, most existing AI systems only process an individual modality. This paper presents an approach that excels at handling multiple modalities of information with a single model. In our "{SkillNet}" model, different parts of the parameters are specialized for processing different modalities. Unlike traditional dense models that always activate all the model parameters, our model sparsely activates parts of the parameters whose skills are relevant to the task. Such model design enables SkillNet to learn skills in a more interpretable way. We develop our model for five modalities including text, image, sound, video and code. Results show that, SkillNet performs comparably to five modality-specific fine-tuned models. Moreover, our model supports self-supervised pretraining with the same sparsely activated way, resulting in better initialized parameters for different modalities. We find that pretraining significantly improves the performance of SkillNet on five modalities, on par with or even better than baselines with modality-specific pretraining. On the task of Chinese text-to-image retrieval, our final system achieves higher accuracy than existing leading systems including Wukong{ViT-B} and Wenlan 2.0 while using less number of activated parameters.

Fan Zhang, Duyu Tang, Yong Dai et al.

Prevailing deep models are single-purpose and overspecialize at individual tasks. However, when being extended to new tasks, they typically forget previously learned skills and learn from scratch. We address this issue by introducing SkillNet-NLU, a general-purpose model that stitches together existing skills to learn new tasks more effectively. The key feature of our approach is that it is sparsely activated guided by predefined skills. Different from traditional dense models that always activate all the model parameters, SkillNet-NLU only activates parts of the model parameters whose skills are relevant to the target task. When learning for a new task, our approach precisely activates required skills and also provides an option to add new skills. We evaluate on natural language understandings tasks and have the following findings. First, with only one model checkpoint, SkillNet-NLU performs better than task-specific fine-tuning and two multi-task learning baselines (i.e., dense model and Mixture-of-Experts model) on six tasks. Second, sparsely activated pre-training further improves the overall performance. Third, SkillNet-NLU significantly outperforms baseline systems when being extended to new tasks.

Shuming Shi, Enbo Zhao, Duyu Tang et al.

In this technical report, we introduce Effidit (Efficient and Intelligent Editing), a digital writing assistant that facilitates users to write higher-quality text more efficiently by using artificial intelligence (AI) technologies. Previous writing assistants typically provide the function of error checking (to detect and correct spelling and grammatical errors) and limited text-rewriting functionality. With the emergence of large-scale neural language models, some systems support automatically completing a sentence or a paragraph. In Effidit, we significantly expand the capacities of a writing assistant by providing functions in five categories: text completion, error checking, text polishing, keywords to sentences (K2S), and cloud input methods (cloud IME). In the text completion category, Effidit supports generation-based sentence completion, retrieval-based sentence completion, and phrase completion. In contrast, many other writing assistants so far only provide one or two of the three functions. For text polishing, we have three functions: (context-aware) phrase polishing, sentence paraphrasing, and sentence expansion, whereas many other writing assistants often support one or two functions in this category. The main contents of this report include major modules of Effidit, methods for implementing these modules, and evaluation results of some key methods.

Cong Zhou, Yong Dai, Duyu Tang et al.

Chinese BERT models achieve remarkable progress in dealing with grammatical errors of word substitution. However, they fail to handle word insertion and deletion because BERT assumes the existence of a word at each position. To address this, we present a simple and effective Chinese pretrained model. The basic idea is to enable the model to determine whether a word exists at a particular position. We achieve this by introducing a special token \texttt{[null]}, the prediction of which stands for the non-existence of a word. In the training stage, we design pretraining tasks such that the model learns to predict \texttt{[null]} and real words jointly given the surrounding context. In the inference stage, the model readily detects whether a word should be inserted or deleted with the standard masked language modeling function. We further create an evaluation dataset to foster research on word insertion and deletion. It includes human-annotated corrections for 7,726 erroneous sentences. Results show that existing Chinese BERT performs poorly on detecting insertion and deletion errors. Our approach significantly improves the F1 scores from 24.1\% to 78.1\% for word insertion and from 26.5\% to 68.5\% for word deletion, respectively.

Jiatong Shi, Jionghao Han, Yichen Lu et al.

Role-play has become a key testbed for generative models, expanding from text-only dialogue to multimodal interaction. Extending role-play to speech captures prosody, emotion, and delivery, but also poses new evaluation challenges. Current pipelines often use audio large language models (ALLMs) as zero-shot judges, which miss paralinguistic cues, collapse multiple aspects into coarse scores, and rely on synthetic speech references that fail to reflect real-world roles. We present Speech-DRAME, a unified framework that contributes at three levels: (i) Speech-DRAME-EvalBench, an evaluation benchmark with bilingual human-annotated data and protocols for training and testing speech evaluation models (SEMs), (ii) DRAME-Eval, a fine-tuned evaluation model, which substantially outperforms zero-shot and few-shot ALLMs, and (iii) Speech-DRAME-RoleBench, a speech role-play benchmark that leverages DRAME-Eval as an automatic judge to compare speech foundation models (SFMs). Speech-DRAME distinguishes between two complementary evaluation strategies: Archetype Evaluation, a top-down approach measuring adherence to broad role archetypes, and Realism Evaluation, a bottom-up approach grounded in real human speech that emphasizes nuanced role quality. Compared to zero-shot ALLM judges, DRAME-Eval achieves stronger agreement with human ratings (Pearson correlation from 0.480 to 0.629 in archetypes, and 0.390 to 0.625 in realism). By integrating transparent benchmark resources, modeling approaches, and system-level evaluation, Speech-DRAME provides the first comprehensive, reproducible foundation for assessing spoken role-play.

Bingyan Xie, Cong Zhou, Yuxuan Shi et al.

This paper proposes a MIMO-OFDM-based context video semantic transmission framework, namely M-CVST, for robust video communication over multi-path multiple-input multiple-output (MIMO) channels. It introduces a context-subcarrier correlation map that aligns video feature context with groups of MIMO subcarriers. To leverage the time-correlated nature of multi-path channels, a recursive subcarrier sampling method paired with time-correlated reference embedding is designed, enabling the use of previously sampled MIMO subcarrier CSI to enhance channel state awareness in the entropy coding model. Numerical results verify the superiority of proposed M-CVST over MIMO multi-path channels compared to other semantic schemes and traditional separated schemes.

Zhiling Ye, Cong Zhou, Xiubao Zhang et al.

In this paper, we explore a reconstruction and reenactment separated framework for 3D Gaussians head, which requires only a single portrait image as input to generate controllable avatar. Specifically, we developed a large-scale one-shot gaussian head generator built upon WebSSL and employed a two-stage training approach that significantly enhances the capabilities of generalization and high-frequency texture reconstruction. During inference, an ultra-lightweight gaussian avatar driven by control signals enables high frame-rate rendering, achieving 90 FPS at a resolution of 512x512. We further demonstrate that the proposed framework follows the scaling law, whereby increasing the parameter scale of the reconstruction module leads to improved performance. Moreover, thanks to the separation design, driving efficiency remains unaffected. Finally, extensive quantitative and qualitative experiments validate that our approach outperforms current state-of-the-art methods.

Biao Guo, Cong Zhou, Fangmin Guo et al.

Most of the current top-down multi-person pose estimation lightweight methods are based on multi-branch parallel pure CNN network architecture, which often struggle to capture the global context required for detecting semantically complex keypoints and are hindered by high latency due to their intricate and redundant structures. In this article, an approximate single-branch lightweight global modeling network (LGM-Pose) is proposed to address these challenges. In the network, a lightweight MobileViM Block is designed with a proposed Lightweight Attentional Representation Module (LARM), which integrates information within and between patches using the Non-Parametric Transformation Operation(NPT-Op) to extract global information. Additionally, a novel Shuffle-Integrated Fusion Module (SFusion) is introduced to effectively integrate multi-scale information, mitigating performance degradation often observed in single-branch structures. Experimental evaluations on the COCO and MPII datasets demonstrate that our approach not only reduces the number of parameters compared to existing mainstream lightweight methods but also achieves superior performance and faster processing speeds.

Zhangyin Feng, Duyu Tang, Cong Zhou et al.

The standard BERT adopts subword-based tokenization, which may break a word into two or more wordpieces (e.g., converting "lossless" to "loss" and "less"). This will bring inconvenience in following situations: (1) what is the best way to obtain the contextual vector of a word that is divided into multiple wordpieces? (2) how to predict a word via cloze test without knowing the number of wordpieces in advance? In this work, we explore the possibility of developing BERT-style pretrained model over a vocabulary of words instead of wordpieces. We call such word-level BERT model as WordBERT. We train models with different vocabulary sizes, initialization configurations and languages. Results show that, compared to standard wordpiece-based BERT, WordBERT makes significant improvements on cloze test and machine reading comprehension. On many other natural language understanding tasks, including POS tagging, chunking and NER, WordBERT consistently performs better than BERT. Model analysis indicates that the major advantage of WordBERT over BERT lies in the understanding for low-frequency words and rare words. Furthermore, since the pipeline is language-independent, we train WordBERT for Chinese language and obtain significant gains on five natural language understanding datasets. Lastly, the analyse on inference speed illustrates WordBERT has comparable time cost to BERT in natural language understanding tasks.

Hao-Wen Dong, Cong Zhou, Taylor Berg-Kirkpatrick et al.

Music performance synthesis aims to synthesize a musical score into a natural performance. In this paper, we borrow recent advances in text-to-speech synthesis and present the Deep Performer -- a novel system for score-to-audio music performance synthesis. Unlike speech, music often contains polyphony and long notes. Hence, we propose two new techniques for handling polyphonic inputs and providing a fine-grained conditioning in a transformer encoder-decoder model. To train our proposed system, we present a new violin dataset consisting of paired recordings and scores along with estimated alignments between them. We show that our proposed model can synthesize music with clear polyphony and harmonic structures. In a listening test, we achieve competitive quality against the baseline model, a conditional generative audio model, in terms of pitch accuracy, timbre and noise level. Moreover, our proposed model significantly outperforms the baseline on an existing piano dataset in overall quality.

Qing Wang, Jia Zhu, Hongji Shu et al.

In natural language processing, extreme multi-label text classification is an emerging but essential task. The problem of extreme multi-label text classification (XMTC) is to recall some of the most relevant labels for a text from an extremely large label set. Large-scale pre-trained models have brought a new trend to this problem. Though the large-scale pre-trained models have made significant achievements on this problem, the valuable fine-tuned methods have yet to be studied. Though label semantics have been introduced in XMTC, the vast semantic gap between texts and labels has yet to gain enough attention. This paper builds a new guide network (GUDN) to help fine-tune the pre-trained model to instruct classification later. Furthermore, GUDN uses raw label semantics combined with a helpful label reinforcement strategy to effectively explore the latent space between texts and labels, narrowing the semantic gap, which can further improve predicted accuracy. Experimental results demonstrate that GUDN outperforms state-of-the-art methods on Eurlex-4k and has competitive results on other popular datasets. In an additional experiment, we investigated the input lengths' influence on the Transformer-based model's accuracy. Our source code is released at https://t.hk.uy/aFSH.

Janusz Klejsa, Per Hedelin, Cong Zhou et al.

We provide a speech coding scheme employing a generative model based on SampleRNN that, while operating at significantly lower bitrates, matches or surpasses the perceptual quality of state-of-the-art classic wide-band codecs. Moreover, it is demonstrated that the proposed scheme can provide a meaningful rate-distortion trade-off without retraining. We evaluate the proposed scheme in a series of listening tests and discuss limitations of the approach.

Cong Zhou, Michael Horgan, Vivek Kumar et al.

Here we present a novel approach to conditioning the SampleRNN generative model for voice conversion (VC). Conventional methods for VC modify the perceived speaker identity by converting between source and target acoustic features. Our approach focuses on preserving voice content and depends on the generative network to learn voice style. We first train a multi-speaker SampleRNN model conditioned on linguistic features, pitch contour, and speaker identity using a multi-speaker speech corpus. Voice-converted speech is generated using linguistic features and pitch contour extracted from the source speaker, and the target speaker identity. We demonstrate that our system is capable of many-to-many voice conversion without requiring parallel data, enabling broad applications. Subjective evaluation demonstrates that our approach outperforms conventional VC methods.