Other people with similar names: Jian Zhu

Unverified author pages with similar names: Jian Zhu


2026

Authorship attribution aims to determine whether a query text and a candidate document are written by the same author. %While dense embeddings are effective, recent work suggests that low-dimensional representations limit encoding capacity. We investigate high-dimensional learned sparse representations (LSR) and multi-vector retrieval for authorship attribution. We further propose Sparse ColBERT (SparCol), a late-interaction sparse lexical reranking model. To address the memory cost of vocabulary-size embeddings, we develop fused Triton kernels that can reduce peak memory usage by 5 11 times over the baseline torch implementation. Extensive experiments show that LSR can outperform dense embeddings in both single-genre and cross-genre settings, whereas improves reranking performance over single-vector models only in single-genre settings due to lexical matching bias. In addition to token-based embeddings, we also explore Lexicon-based embeddings that score clusters of tokens, which is competitive with the vocabulary-based LSR with reduced dimensions. We conduct interpretability analysis of LSRs, showing that they allow for a more transparent scoring process.
We perform in-depth evaluations of in-context learning (ICL) on state-of-the-art transformer, state-space, and hybrid large language models over two categories of knowledge-based ICL tasks. Using a combination of behavioral probing and intervention-based methods, we have discovered that, while LLMs of different architectures can behave similarly in task performance, their internals could remain different. We discover that function vectors (FVs) responsible for ICL are primarily located in the self-attention and Mamba layers, and speculate that Mamba2 uses a different mechanism from FVs to perform ICL. FVs are more important for ICL involving parametric knowledge retrieval, but not for contextual knowledge understanding. Our work contributes to a more nuanced understanding across architectures and task types. Methodologically, our approach also highlights the importance of combining both behavioural and mechanistic analyses to investigate LLM capabilities.
We introduce OntologyBench, a tiered biomedical retrieval benchmark comprising 471,854 training and 125,744 evaluation query–document relevance pairs across concept grounding, relational retrieval, and compositional phenotype-based retrieval.

Although these tasks can be tractable using ontology-aware reference methods, across task tiers, embedding performance is generally lower on relational and compositional tasks than on concept-grounding tasks. Fine-tuning on ontology-derived supervision improves performance on several relational and compositional tasks, whereas the evaluated reranking and LLM-based candidate-scoring methods provide little or no end-to-end improvement. Errors frequently reflect diseases matching only subsets of the phenotype evidence.

These findings indicate that the evaluated embedding and reranking configurations do not reliably recover the compatibility encoded by the selected ontology relations and phenotype combinations and motivate retrieval systems that better integrate learned representations with structured biomedical knowledge.
Linguistic typology relies on expert analysis of reference grammars across languages, making large-scale crosslinguistic comparison labor-intensive and unscalable. We introduce AutoTypologist, an LLM agent for evidence-grounded typological analysis over reference grammars. The agent is capable of retrieving relevant grammar sections, analyzing interlinear glossed text (IGT), and iteratively reasoning over typological hypotheses using a ReAct-style workflow. We evaluate the system on Typological Feature Coding against expert annotations and Typological Hypothesis Testing with typological universals using 25 open-source reference grammars. Operating under different information constraints in Typological Feature Coding, the agent can synthesize information from reference grammar prose but still faces challenges with only IGTs in the target language. In Typological Hypothesis Testing, the agent can synthesize crosslinguistic evidence and identify both supporting cases and counterexamples. These findings suggest that LLM agents can support scalable and inspectable typological analysis, while still requiring expert validation.
Recent work has shown that LLMs overrepresent dominant cultures, particularly Western ones, while marginalizing others. We investigate whether this affects models’ ability to generate culturally adapted responses by evaluating their use of local measurement units based on the user’s perceived cultural background. We introduce Cultural and Pragmatic Response Inference (CAPRI), a dataset of conversations with varying levels of cultural cues. Experiments with state-of-the-art LLMs show that models can infer cultural background and recall relevant conventions, but often fail to utilize the information to adapt their answers to the relevant cultural conventions, unless explicitly prompted to perform the tasks sequentially. We further evaluate adaptation to the interpretation of time and quantity expressions, two subjective language grounding dimensions that are affected by culture. We find that models increasingly adapt their answers as cultural cues accumulate, but their priors are not culture-neutral, sometimes aligning with the model’s country of origin. Overall, CAPRI provides a resource for future research aimed at narrowing the gap between cultural knowledge and culturally adaptive language generation.
Linguistic style is central to how texts convey meaning and to tasks from register analysis to authorship verification, yet extracting detailed stylistic features at scale is hard: rule-based taggers are interpretable but slow, while neural style embeddings are fast but opaque. We present Neurobiber, a RoBERTa tagger that predicts 96 interpretable features from Biber’s Multidimensional Analysis in a single forward pass, bringing this analysis to corpus scale: a one-billion-token corpus tags on a single GPU in about 9.5 hours, over an order of magnitude faster than single-thread taggers. Neurobiber is trained on BiberPlus, a new open-source rule tagger that computes the same 96 features deterministically from explicit rules and ships built-in analytics such as PCA and factor analysis; users tag at scale with Neurobiber and fall back to BiberPlus for exact, auditable counts. Neurobiber reproduces BiberPlus at 0.97 macro-F1; its features recover Biber’s primary dimension of variation, and on PAN 2020 authorship verification they contribute signal complementary to a strong neural encoder. Both tools are released openly, with a live demo.
Phone recognition (PR) serves as the atomic interface for language-agnostic modeling for cross-lingual speech processing and phonetic analysis. Despite prolonged efforts in developing PR systems, current evaluations only measure surface-level transcription accuracy. We introduce PRiSM, the first open-source benchmark designed to expose blind spots in phonetic perception through intrinsic and extrinsic evaluation of PR systems. PRiSM standardizes transcription-based evaluation and assesses downstream utility in clinical, educational, and multilingual settings with transcription and representation probes. We find that diverse language exposure during training is key to PR performance, encoder-CTC models are the most stable, and specialized PR systems still outperform LALMs. PRiSM releases code, recipes, and datasets to move the field toward multilingual speech models with robust phonetic ability.
Recent advances in spoken language processing have led to substantial progress in phonetic tasks such as automatic speech recognition (ASR), phone recognition (PR), grapheme-to-phoneme conversion (G2P), and phoneme-to-grapheme conversion (P2G). Despite their conceptual similarity, these tasks have largely been studied in isolation, each relying on task-specific architectures and datasets. In this paper, we introduce POWSM (Phonetic Open Whisper-style Speech Model), the first unified framework capable of jointly performing multiple phone-related tasks. POWSM enables seamless conversion between audio, text (graphemes), and phones, opening up new possibilities for universal and low-resource speech processing. Our model outperforms or matches specialized PR models of similar size (Wav2Vec2Phoneme and ZIPA) while jointly supporting G2P, P2G, and ASR. Our training data, models, and code are released to foster open science.

2025

Curating datasets that span multiple languages is challenging. To make the collection more scalable, researchers often incorporate one or more imperfect classifiers in the process, like language identification models. These models, however, are prone to failure, resulting in some language partitions being unreliable for downstream tasks. We introduce a statistical test, the Preference Proportion Test, for identifying such unreliable partitions. By annotating only 20 samples for a language partition, we are able to identify systematic transcription errors for 10 language partitions in a recent large multilingual transcribed audio archive, X-IPAPack (Zhu et al., 2024). We find that filtering these low-quality partitions out when training models for the downstream task of phonetic transcription brings substantial benefits, most notably a 25.7% relative improvement on transcribing recordings in out-of-distribution languages. Our work contributes an effective method for auditing multilingual audio archives.1
We present lightweight flow matching multilingual text-to-speech (TTS) systems for Ojibwe, Mi’kmaq, and Maliseet, three Indigenous languages in North America. Our results show that training a multilingual TTS model on three typologically similar languages can improve the performance over monolingual models, especially when data are scarce. Attention-free architectures are highly competitive with self-attention architecture with higher memory efficiency. Our research provides technical development to language revitalization for low-resource languages but also highlights the cultural gap in human evaluation protocols, calling for a more community-centered approach to human evaluation.
This paper introduces LingGym, a new benchmark that evaluates LLMs’ capacity for meta-linguistic reasoning using Interlinear Glossed Text (IGT) and grammatical descriptions extracted from 18 typologically diverse reference grammars. Unlike previous work that focuses on specific downstream tasks, we assess whether LLMs can generalize linguistic inference across low-resource languages and structures not seen during training. We present a controlled evaluation task: Word-Gloss Inference, in which the model must infer a missing word and gloss from context using varying levels of linguistic information (e.g., glosses, grammatical explanations, translations). Our results show that incorporating structured linguistic cues leads to consistent improvements in reasoning performance across all models. This work highlights both the promise and current limitations of using LLMs for typologically informed linguistic analysis and low-resource language documentation.
Case reports provide critical insights into rare and atypical diseases, but extracting structured knowledge remains challenging due to unstructured text and domain-specific terminology. We introduce CaseReportCollective, an LLM-extracted dataset of 85,961 open-access case reports spanning 37 years across 14 medical domains, validated through programmatic and human evaluation. Our dataset reveals key publication and demographic trends, including a significant increase in open-access case reports over the past decade, shifts in focus from oncology to COVID-19, and sex disparities in reporting across different medical conditions. Over time, the gap between male and female case reports has narrowed, suggesting greater equity in case reporting. Using CaseReportCollective, we further explore embedding-based retrieval for similar medical topics through accumulated similarity scores across extracted structured information. We also conducted detailed error analyses on the retrieval ranking, finding that high-reported topics dominate retrieval. Such retrieval is driven by lexical overlap rather than underlying clinical relevance, often failing to distinguish between semantically similar yet mechanistically distinct conditions. Future work should focus on clinical-aware embeddings adjusted for long-tailed case distributions to improve retrieval accuracy.
We present ZIPA, a family of efficient speech models that advances the state-of-the-art performance of crosslinguistic phone recognition. We first curated IPA PACK++, a large-scale multilingual speech corpus with 17,000+ hours of normalized phone transcriptions and a novel evaluation set capturing unseen languages and sociophonetic variation. ZIPA, including transducer (ZIPA-T) and CTC-based (ZIPA-CR) variants, leverages the efficient Zipformer backbones and outperforms existing phone recognition systems with much fewer parameters. Further scaling via noisy student training on 11,000+ hours of pseudo-labeled multilingual data yields further improvement. While ZIPA achieves strong performance on benchmarks, error analysis reveals persistent limitations in modeling sociophonetic diversity, underscoring challenges for future research.

2024

In this project, we demonstrate that phoneme-based models for speech processing can achieve strong crosslinguistic generalizability to unseen languages. We curated the IPAPACK, a massively multilingual speech corpora with phonemic transcriptions, encompassing more than 115 languages from diverse language families, selectively checked by linguists. Based on the IPAPACK, we propose CLAP-IPA, a multi-lingual phoneme-speech contrastive embedding model capable of open-vocabulary matching between arbitrary speech signals and phonemic sequences. The proposed model was tested on 95 unseen languages, showing strong generalizability across languages. Temporal alignments between phonemes and speech signals also emerged from contrastive training, enabling zeroshot forced alignment in unseen languages. We further introduced a neural forced aligner IPA-ALIGNER by finetuning CLAP-IPA with the Forward-Sum loss to learn better phone-to-audio alignment. Evaluation results suggest that IPA-ALIGNER can generalize to unseen languages without adaptation.

2023

Transformers have revolutionized almost all natural language processing (NLP) tasks but suffer from memory and computational complexity that scales quadratically with sequence length. In contrast, recurrent neural networks (RNNs) exhibit linear scaling in memory and computational requirements but struggle to match the same performance as Transformers due to limitations in parallelization and scalability. We propose a novel model architecture, Receptance Weighted Key Value (RWKV), that combines the efficient parallelizable training of transformers with the efficient inference of RNNs. Our approach leverages a linear attention mechanism and allows us to formulate the model as either a Transformer or an RNN, thus parallelizing computations during training and maintains constant computational and memory complexity during inference. We scale our models as large as 14 billion parameters, by far the largest dense RNN ever trained, and find RWKV performs on par with similarly sized Transformers, suggesting future work can leverage this architecture to create more efficient models. This work presents a significant step towards reconciling trade-offs between computational efficiency and model performance in sequence processing tasks.

2022

Inducing semantic representations directly from speech signals is a highly challenging task but has many useful applications in speech mining and spoken language understanding. This study tackles the unsupervised learning of semantic representations for spoken utterances. Through converting speech signals into hidden units generated from acoustic unit discovery, we propose WavEmbed, a multimodal sequential autoencoder that predicts hidden units from a dense representation of speech. Secondly, we also propose S-HuBERT to induce meaning through knowledge distillation, in which a sentence embedding model is first trained on hidden units and passes its knowledge to a speech encoder through contrastive learning. The best performing model achieves a moderate correlation (0.5 0.6) with human judgments, without relying on any labels or transcriptions. Furthermore, these models can also be easily extended to leverage textual transcriptions of speech to learn much better speech embeddings that are strongly correlated with human annotations. Our proposed methods are applicable to the development of purely data-driven systems for speech mining, indexing and search.

2021

New words are regularly introduced to communities, yet not all of these words persist in a community’s lexicon. Among the many factors contributing to lexical change, we focus on the understudied effect of social networks. We conduct a large-scale analysis of over 80k neologisms in 4420 online communities across a decade. Using Poisson regression and survival analysis, our study demonstrates that the community’s network structure plays a significant role in lexical change. Apart from overall size, properties including dense connections, the lack of local clusters, and more external contacts promote lexical innovation and retention. Unlike offline communities, these topic-based communities do not experience strong lexical leveling despite increased contact but accommodate more niche words. Our work provides support for the sociolinguistic hypothesis that lexical change is partially shaped by the structure of the underlying network but also uncovers findings specific to online communities.
An individual’s variation in writing style is often a function of both social and personal attributes. While structured social variation has been extensively studied, e.g., gender based variation, far less is known about how to characterize individual styles due to their idiosyncratic nature. We introduce a new approach to studying idiolects through a massive cross-author comparison to identify and encode stylistic features. The neural model achieves strong performance at authorship identification on short texts and through an analogy-based probing task, showing that the learned representations exhibit surprising regularities that encode qualitative and quantitative shifts of idiolectal styles. Through text perturbation, we quantify the relative contributions of different linguistic elements to idiolectal variation. Furthermore, we provide a description of idiolects through measuring inter- and intra-author variation, showing that variation in idiolects is often distinctive yet consistent.

2019

This paper describes the UM-IU@LING’s system for the SemEval 2019 Task 6: Offens-Eval. We take a mixed approach to identify and categorize hate speech in social media. In subtask A, we fine-tuned a BERT based classifier to detect abusive content in tweets, achieving a macro F1 score of 0.8136 on the test data, thus reaching the 3rd rank out of 103 submissions. In subtasks B and C, we used a linear SVM with selected character n-gram features. For subtask C, our system could identify the target of abuse with a macro F1 score of 0.5243, ranking it 27th out of 65 submissions.