VAE

A Persona-Aware LLM-Enhanced Framework for Multi-Session Personalized Dialogue Generation (2025.findings-acl)

Challenge:	TextBox is an open-source text generation framework that is modularized and extensible.
Approach:	They propose to provide a unified, modularized, and extensible text generation framework that implements 21 text generation models on 9 benchmark datasets.
Outcome:	The proposed framework implements 21 models on 9 benchmark datasets and is available under the Apache License 2.0 license.

Copied to clipboard

Dongshuo Liu, Zhijing Wu, Dandan Song, Heyan Huang

On the Importance of the Kullback-Leibler Divergence Term in Variational Autoencoders for Text Generation (D19-56)

Challenge:	Existing personalized dialogue models focus on dialogue history and personality information, reducing the responses’ consistency.
Approach:	They propose a Persona-Aware LLM-enAnCEd(PALACE) framework that generates responses consistent with dialogue history and personality information across multiple sessions to engage users’ interest in the dialogue.
Outcome:	The proposed framework outperforms the state-of-the-art methods in automatic and human evaluation metrics on the MSC and DuLeMon datasets.

Copied to clipboard

Victor Prokhorov, Ehsan Shareghi, Yingzhen Li, Mohammad Taher Pilehvar, Nigel Collier

Topic-Guided Variational Auto-Encoder for Text Generation (N19-1)

Challenge:	Variational Autoencoders suffer from learning uninformative latent representations due to issues such as approximated posterior collapse or entanglement of the latent space.
Approach:	They propose to impose an explicit constraint on the Kullback-Leibler divergence term inside the VAE objective function to understand the significance of the KL term in controlling the information transmitted through the VAe channel.
Outcome:	The proposed constraint avoids posterior collapse, but it also controls the information transmitted through the VAE channel.

Copied to clipboard

Wenlin Wang, Zhe Gan, Hongteng Xu, Ruiyi Zhang, Guoyin Wang, Dinghan Shen, Changyou Chen, Lawrence Carin

Document Hashing with Multi-Grained Prototype-Induced Hierarchical Generative Model (2024.findings-emnlp)

Challenge:	Experimental results show that our model outperforms its competitors on both unconditional and conditional text generation.
Approach:	They propose a topic-guided variational auto-encoder model for text generation that specifies a Gaussian mixture model and a neural topic module to generate sentences under the topic.
Outcome:	The proposed model outperforms existing variational auto-encoders on unconditional and conditional text generation, and can generate semantically-meaningful sentences with various topics.

Copied to clipboard

Qian Zhang, Qinliang Su, Jiayang Chen, Zhenpeng Song

Challenge:	Existing document hashing methods only consider flat semantics of documents, preserving hierarchical semantics.
Approach:	They propose a hierarchical generative model that can model and leverage hierarchic semantics . they introduce hierarchically-based prototypes into the model to construct a Hierarchical prior distribution .
Outcome:	The proposed model outperforms baseline methods on hierarchical and flat datasets.

Cyclical Annealing Schedule: A Simple Approach to Mitigating KL Vanishing (N19-1)

Copied to clipboard

Hao Fu, Chunyuan Li, Xiaodong Liu, Jianfeng Gao, Asli Celikyilmaz, Lawrence Carin

VOLTA: Improving Generative Diversity by Variational Mutual Information Maximizing Autoencoder (2024.findings-naacl)

Challenge:	Variational autoencoders (VAEs) with an auto-regressive decoder have been applied for many natural language processing tasks.
Approach:	They propose a cyclical annealing schedule which repeats the process of increasing multiple times to learn more meaningful latent codes progressively by leveraging previous learning cycles as warm re-restart.
Outcome:	The proposed method improves on a broad range of NLP tasks, including language modeling, dialog response generation and semi-supervised text classification.

Copied to clipboard

Yueen Ma, DaFeng Chi, Jingjing Li, Kai Song, Yuzheng Zhuang, Irwin King

Cross-Utterance Conditioned VAE for Non-Autoregressive Text-to-Speech (2022.acl-long)

Challenge:	generative diversity is a critical yet underexplored issue in natural language generation . previous approaches to enhance diversity of Transformer models have been limited by their latent variables .
Approach:	They propose a framework that bridges Transformer with VAE to enhance generative diversity.
Outcome:	The proposed framework improves generative diversity while maintaining generative quality.

Copied to clipboard

Yang Li, Cheng Yu, Guangzhi Sun, Hua Jiang, Fanglei Sun, Weiqin Zu, Ying Wen, Yang Yang, Jun Wang

Multimodality

Challenge:	Experimental results show that the proposed model improves naturalness and prosody diversity with clear margins.
Approach:	They propose a cross-utterance conditional VAE to estimate posterior probability distribution of latent prosody features for each phoneme by conditioning on acoustic features, speaker information, and text features from past and future sentences.
Outcome:	The proposed model improves naturalness and prosody diversity with clear margins.

Graph-Induced Syntactic-Semantic Spaces in Transformer-Based Variational AutoEncoders (2024.findings-naacl)

Copied to clipboard

Yingji Zhang, Marco Valentino, Danilo Carvalho, Ian Pratt-Hartmann, Andre Freitas

Polarized-VAE: Proximity Based Disentangled Representation Learning for Text Generation (2021.eacl-main)

Challenge:	Existing studies on syntactic injection in Variational AutoEncoders (VAEs) are limited to LSTM-based VAEs.
Approach:	They propose to use latent space separation techniques to inject syntactic information into Variational AutoEncoders (VAEs) using graph-based models.
Outcome:	The proposed end-to-end VAE architecture can improve the organisation of the latent space, alleviating the information loss occurring in standard VAE setups, and resulting in enhanced performances on language modelling and downstream generation tasks.

Copied to clipboard

Vikash Balasubramanian, Ivan Kobyzev, Hareesh Bahuleyan, Ilya Shapiro, Olga Vechtomova

APo-VAE: Text Generation in Hyperbolic Space (2021.naacl-main)

Challenge:	Existing methods for learning disentangled representations of real-world data focus on attribute labels or unsupervised methods that manipulate factorization in the latent space of models such as the variational autoencoder (VAE).
Approach:	They propose an approach that disentangles select attributes in the latent space based on proximity measures reflecting the similarity between data points with respect to these attributes.
Outcome:	The proposed method outperforms the VAE baseline and is competitive with state-of-the-art approaches while being more a general framework applicable to other attribute disentanglement tasks.

Copied to clipboard

Shuyang Dai, Zhe Gan, Yu Cheng, Chenyang Tao, Lawrence Carin, Jingjing Liu

Robustness in NLP

Challenge:	Existing models that learn embeddings only in Euclidean vector space do not account for such structural property of language.
Approach:	They propose a Poincare Variational Autoencoder to capture latent hierarchies in hyperbolic space . they propose enabling adversarial learning procedures to empower robust model training .
Outcome:	The proposed model outperforms existing models in a hyperbolic latent space . it captures latent language hierarchies in hyperbolical space and is robust to training .

Variational Autoencoder with Embedded Student-t Mixture Model for Authorship Attribution (2020.coling-main)

Copied to clipboard

Benedikt Boenninghoff, Steffen Zeiler, Robert Nickel, Dorothea Kolossa

Non-Autoregressive Neural Machine Translation with Consistency Regularization Optimized Variational Framework (2022.naacl-main)

Challenge:	Existing Variational autoencoders are limited by the assumed Gaussianity of the underlying probability distributions in the latent space.
Approach:	They propose a probabilistic autoencoding framework to deal with a supervised authorship attribution task.
Outcome:	The proposed method outperforms existing methods on an Amazon review dataset.

Copied to clipboard

Minghao Zhu, Junli Wang, Chungang Yan

Machine Translation

Challenge:	Variational Autoencoder (VAE) is an effective framework to model the interdependency for non-autoregressive neural machine translation (NAT).
Approach:	They propose to use Variational Autoencoder to model interdependency for non-autoregressive neural machine translation (NAT) a posterior consistency regularization approach is proposed to improve translation quality .
Outcome:	The proposed model is 1.5/0.7 and 0.8/0.3 BLEU points faster than the baseline model.

Fuse It More Deeply! A Variational Transformer with Layer-Wise Latent Variable Inference for Text Generation (2022.naacl-main)

Copied to clipboard

Jinyi Hu, Xiaoyuan Yi, Wenhao Li, Maosong Sun, Xing Xie

Combining Sentiment Lexica with a Multi-View Variational Autoencoder (N19-1)

Challenge:	Variational Auto-Encoders are often used for text generation tasks due to the sequential nature of the text.
Approach:	They propose a variational Transformer framework that learns a series of layer-wise latent variables with each inferred from those of lower layers and tightly coupled with the hidden states by low-rank tensor product.
Outcome:	The proposed framework can learn latent variables from lower layers and incorporate more information.

Copied to clipboard

Alexander Miserlis Hoyle, Lawrence Wolf-Sonkin, Hanna Wallach, Ryan Cotterell, Isabelle Augenstein

On the Effect of Isotropy on VAE Representations of Text (2022.acl-short)

Challenge:	a new model of sentiment lexica is being developed to combine disparate scales into a common representation.
Approach:	They propose a model that unifies disparate scales into a common latent representation . they evaluate a text classification task using nine English-Language sentiment datasets .
Outcome:	The proposed model outperforms six individual sentiment lexica and a simple combination thereof.

Copied to clipboard

Lan Zhang, Wray Buntine, Ehsan Shareghi

Neural Attention-Aware Hierarchical Topic Model (2021.emnlp-main)

Challenge:	Injecting desired geometric properties into text representations has attracted a lot of attention due to its better utilisation of representation space.
Approach:	They propose to use an isotropic Gaussian posterior instead of the ellipsoidal Gausssian priori to inject isotropy into text representations.
Outcome:	The proposed method improves classification performance, robustness to input perturbation, and generative behavior compared to the ellipsoidal Gaussian posterior.

Copied to clipboard

Yuan Jin, He Zhao, Ming Liu, Lan Du, Wray Buntine

Improving Semantic Control in Discrete Latent Spaces with Transformer Quantized Variational Autoencoders (2024.findings-eacl)

Challenge:	Neural topic models (NTMs) use deep neural networks to learn topic information.
Approach:	They propose a variational autoencoder model that reconstructs sentence and document word counts using bag-of-words embeddings and pre-trained semantic embedders.
Outcome:	The proposed model lowers reconstruction errors at sentence and document levels and finds more coherent topics from real-world datasets.

Copied to clipboard

Yingji Zhang, Danilo Carvalho, Marco Valentino, Ian Pratt-Hartmann, Andre Freitas

Generative Text Modeling through Short Run Inference (2021.eacl-main)

Challenge:	Recent work has struggled to achieve consistent results due to the inevitable loss of semantic information in the variational bottleneck and limited control over the decoding mechanism.
Approach:	They propose a model that leverages the controllability of VQVAE to guide the self-attention mechanism in Transformer-based VAEs to improve semantic control and generation.
Outcome:	The proposed model outperforms existing state-of-the-art VAE models in terms of control and preservation of semantic information across different tasks.

Copied to clipboard

Bo Pang, Erik Nijkamp, Tian Han, Ying Nian Wu

Challenge:	Latent variable models for text capture global semantic and syntactic features when trained correctly.
Approach:	They propose a short run dynamics for inference that initializes from the prior distribution of the latent variable and runs a small number of Langevin dynamics steps guided by its posterior distribution.
Outcome:	The proposed model is able to generate coherent sentences with smooth transition and shows no sign of posterior collapse.

Learning VAE-LDA Models with Rounded Reparameterization Trick (2020.emnlp-main)

Copied to clipboard

Runzhi Tian, Yongyi Mao, Richong Zhang

BERT-Flow-VAE: A Weakly-supervised Model for Multi-Label Text Classification (2022.coling-1)

Challenge:	No reparameterization form of Dirichlet distributions is known to date for topic models .
Approach:	They propose a method to reparameterize Dirichlet distributions for the learning of VAE-LDA models by using a latent Dirichlets prior.
Outcome:	The proposed method outperforms existing neural topic models on benchmark datasets and on a synthetic dataset.

Copied to clipboard

Ziwen Liu, Josep Grau-Bove, Scott Allan Orr

CVAE-based Re-anchoring for Implicit Discourse Relation Classification (2021.findings-emnlp)

Challenge:	Multi-label Text Classification (MLTC) is a task of categorizing documents into one or more topics. Fully-supervised learning methods are undesirable for this task because of the diversity of domains of application and cost of manual labelling.
Approach:	They propose a Weakly-Supervised Multi-Label Text Classification model that produces BERT sentence embeddings and calibrates them using a flow model.
Outcome:	The proposed model outperforms baseline models in key metrics and achieves 84% performance on multi-label datasets.

Copied to clipboard

Zujun Dou, Yu Hong, Yu Sun, Guodong Zhou

Text Classification

Challenge:	Existing studies show that training implicit discourse relation classifiers suffers from data sparsity.
Approach:	They propose a re-anchoring strategy to reduce the risk of erroneous sampling . they use Conditional VAE to estimate the risk and migrate the anchor to reduce it .
Outcome:	The proposed method improves the baseline classifier performance on PDTB v2.0 .

Learning Disentangled Semantic Spaces of Explanations via Invertible Neural Networks (2024.acl-long)

Copied to clipboard

Yingji Zhang, Danilo Carvalho, Andre Freitas

Explainability & Interpretability in NLP

Challenge:	Disentangled latent spaces usually have better semantic separability and geometrical properties, which leads to better interpretability and controllable data generation.
Approach:	They propose a flow-based invertible neural network mechanism integrated with a transformer-based language Autoencoder to deliver latent spaces with better semantic separability and geometrical properties.
Outcome:	The proposed model can deliver latent spaces with better separability properties compared to the current state-of-the-art models.

Diversity-Aware Coherence Loss for Improving Neural Topic Models (2023.acl-short)

Copied to clipboard

Raymond Li, Felipe Gonzalez-Pizarro, Linzi Xing, Gabriel Murray, Giuseppe Carenini

Unsupervised Relation Extraction: A Variational Autoencoder Approach (2021.emnlp-main)

Challenge:	Experimental results show that our method significantly improves the performance of neural topic models without requiring any pretraining or additional parameters.
Approach:	They propose a variational autoencoder framework that minimizes the posterior and prior divergence and a diversity-aware coherence loss that encourages the model to learn corpus-level coherency scores while maintaining high diversity between topics.
Outcome:	The proposed approach significantly improves the performance of neural topic models without pretraining or additional parameters.

Copied to clipboard

Chenhan Yuan, Hoda Eldardiry

Style Pooling: Automatic Text Style Obfuscation for Improved Classification Fairness (2021.emnlp-main)

Challenge:	Existing methods for relation extraction use latent variables and supervised training which requires large datasets.
Approach:	They propose a VAE-based unsupervised relation extraction technique that uses latent variables as an intermediate variable instead of a latent variable.
Outcome:	The proposed method outperforms state-of-the-art methods on the NYT dataset and outperformed existing methods.

Copied to clipboard

Fatemehsadat Mireshghallah, Taylor Berg-Kirkpatrick

Text Classification

Challenge:	Text style can reveal sensitive attributes of the author (e.g. age and race) to the reader, which can lead to privacy violations and bias in both human and algorithmic decisions based on text.
Approach:	They propose a framework that obfuscates stylistic features of human-generated text through style transfer by automatically re-writing the text itself.
Outcome:	The proposed framework obfuscates stylistic features of human-generated text through style transfer, by automatically re-writing the text itself.

A Hierarchical Latent Structure for Variational Conversation Modeling (N18-1)

Copied to clipboard

Yookoon Park, Jaemin Cho, Gunhee Kim

Do Neural Topic Models Really Need Dropout? Analysis of the Effect of Dropout in Topic Modeling (2023.eacl-main)

Challenge:	Variational autoencoders suffer from the notorious degeneration problem, according to a new study . utterance drop regularization is an important feature of the hierarchical RNNs .
Approach:	They propose a variational hierarchical conversation RNN framework that exploits latent variables and an utterance drop regularization to exploit latent variable.
Outcome:	The proposed model outperforms state-of-the-art models on Cornell Movie Dialog and Ubuntu Dialog Corpus.

Copied to clipboard

Suman Adhya, Avishek Lahiri, Debarshi Kumar Sanyal

RegaVAE: A Retrieval-Augmented Gaussian Mixture Variational Auto-Encoder for Language Modeling (2023.findings-emnlp)

Challenge:	Dropout is a regularization trick used to resolve overfitting in large feedforward neural networks, but there is nil analysis of it for unsupervised models and in particular, VAE-based neural topic models.
Approach:	They propose to use dropout to solve overfitting problems in unsupervised neural topic models by stochastically dropping out the activation of neurons to prevent complex co-adaptations of feature vectors.
Outcome:	The proposed class of neural topic models can be used to improve the quality and predictive performance of the generated topics.

Copied to clipboard

Jingcheng Deng, Liang Pang, Huawei Shen, Xueqi Cheng

Unsupervised Dialog Structure Learning (N19-1)

Challenge:	Existing research on retrieval-augmented language models has two main problems: determining what information to retrieve and effectively combining retrieved information during generation.
Approach:	They propose a retrieval-augmented language model that captures current and future information from source and target text into a latent space.
Outcome:	The proposed model is more efficient than explicit raw text, but limited by context length and noise.

Copied to clipboard

Weiyan Shi, Tiancheng Zhao, Zhou Yu

T2S-GPT: Dynamic Vector Quantization for Autoregressive Sign Language Production from Text (2024.acl-long)

Challenge:	Current dialog systems require human experts to design the dialog structure, which is time consuming and sometimes insufficient to satisfy various customer needs.
Approach:	They propose to extract dialog structure using a modified VRNN model with discrete latent vectors.
Outcome:	The proposed model outperforms existing models on the ability to predict unseen data and is faster and more effective in a reinforcement learning setting.

Copied to clipboard

Aoxiong Yin, Haoyuan Li, Kai Shen, Siliang Tang, Yueting Zhuang

Syntax-Infused Variational Autoencoder for Text Generation (P19-1)

Challenge:	Existing vector quantization methods are fixed-length encodings, overlooking the uneven information density in sign language.
Approach:	They propose a two-stage sign language production paradigm that encodes sign language sequences into discrete codes and autoregressively generates sign languages from text.
Outcome:	The proposed model can dynamically adjust the encoding length based on the information density in sign language to achieve accurate and compact encoded enccoding.

Copied to clipboard

Xinyuan Zhang, Yi Yang, Siyang Yuan, Dinghan Shen, Lawrence Carin

Syntactic Text Processing

Challenge:	Experimental results demonstrate the generative superiority of SIVAE on both reconstruction and targeted syntactic evaluations.
Approach:	They propose a syntax-infused variational autoencoder that integrates sentences with their syntactic trees to improve the grammar of generated sentences.
Outcome:	The proposed model improves the grammar of generated sentences by integrating sentences with syntactic trees.

Towards Generating Long and Coherent Text with Multi-Level Latent Variable Models (P19-1)

Copied to clipboard

Dinghan Shen, Asli Celikyilmaz, Yizhe Zhang, Liqun Chen, Xin Wang, Jianfeng Gao, Lawrence Carin

Syntactically-Informed Unsupervised Paraphrasing with Non-Parallel Data (2021.emnlp-main)

Challenge:	Variational autoencoders (VAEs) have received much attention as an end-to-end architecture for text generation with latent variables.
Approach:	They propose to leverage several multi-level structures to learn a variational autoencoder model for generating long, and coherent text.
Outcome:	The proposed model produces more coherent and less repetitive long text compared to baselines and mitigates posterior collapse issue.

Copied to clipboard

Erguang Yang, Mingtong Liu, Deyi Xiong, Yujie Zhang, Yao Meng, Changjian Hu, Jinan Xu, Yufeng Chen

Responsible & Trustworthy NLP

Challenge:	Existing studies on syntactically controlled paraphrase generation rely on large-scale parallel data.
Approach:	They propose a syntactically-informed unsupervised paraphrasing model based on conditional variational auto-encoder which can generate texts in a specified syntastic structure.
Outcome:	The proposed model can generate diverse paraphrases with specified syntactic structure using non-parallel data.

Improving Variational Autoencoder for Text Modelling with Timestep-Wise Regularisation (2020.coling-main)

Copied to clipboard

Ruizhe Li, Xiao Li, Guanyi Chen, Chenghua Lin

Doc2hash: Learning Discrete Latent variables for Documents Retrieval (N19-1)

Challenge:	Variational Autoencoders (VAEs) have been widely used in text modelling but posterior collapse is a problem when RNN-based models are employed.
Approach:	They propose a timestep-wise regularisation VAE architecture which can effectively avoid posterior collapse when used in text modelling.
Outcome:	The proposed model avoids posterior collapse and can be applied to any RNN-based VAE model.

Copied to clipboard

Yifei Zhang, Hao Zhu

Information Retrieval

Challenge:	Learning to hash via generative model is a powerful paradigm for fast similarity search in documents retrieval.
Approach:	They propose a method that trains a generative model to generate hash codes by using continuous relaxation on priors.
Outcome:	The proposed method outperforms other state-of-the-art methods in qualitative and quantitative experiments.

A Batch Normalized Inference Network Keeps the KL Vanishing Away (2020.acl-main)

Copied to clipboard

Qile Zhu, Wei Bi, Xiaojiang Liu, Xiyao Ma, Xiaolin Li, Dapeng Wu

Challenge:	Variational Autoencoder (VAE) is widely used to approximate a model’s posterior on latent variables.
Approach:	They propose to let the Kullback–Leibler divergence individual follow a distribution across the whole dataset and analyze that it is sufficient to prevent posterior collapse by keeping the expectation of the KL’s distribution positive.
Outcome:	The proposed approach can avoid posterior collapse effectively and efficiently without introducing any new model component or modifying the objective.

FlowPrior: Learning Expressive Priors for Latent Variable Sentence Models (2021.naacl-main)

Copied to clipboard

Xiaoan Ding, Kevin Gimpel

Challenge:	Existing training strategies are not effective for learning rich priors, we propose adding the importance-sampled log marginal likelihood as a second term to the standard VAE objective.
Approach:	They propose to add importance-sampled log marginal likelihood to standard VAE objective to help when learning the prior.
Outcome:	The proposed model improves language modeling tasks compared to baselines.

A Probabilistic Generative Model for Typographical Analysis of Early Modern Printing (2020.acl-main)

Copied to clipboard

Kartik Goyal, Chris Dyer, Christopher Warren, Maxwell G’Sell, Taylor Berg-Kirkpatrick

Responsible & Trustworthy NLP

Challenge:	Scholars often need to go beyond textual analysis for establishing provenance of historical documents.
Approach:	They propose a deep and interpretable probabilistic generative model to analyze glyph shapes in printed Early Modern documents by generating a latent vector responsible for inking variations, jitter, noise and other unforeseen phenomena.
Outcome:	The proposed model outperforms interpretable clustering baselines and overly-flexible deep generative models on the task of completely unsupervised discovery of typefaces in mixed-fonts documents.

MacLaSa: Multi-Aspect Controllable Text Generation via Efficient Sampling from Compact Latent Space (2023.findings-emnlp)

Copied to clipboard

Hanxing Ding, Liang Pang, Zihao Wei, Huawei Shen, Xueqi Cheng, Tat-Seng Chua

Green & Sustainable NLP

Challenge:	Existing approaches to multi-aspect controllable text generation require expensive iteration / searching within the discrete text space during the decoding stage, resulting in a degradation of text quality due to the domain discrepancies between different aspects.
Approach:	They propose a framework that estimates compact latent space for multiple aspects and performs efficient Sampling with a fast sampler to eliminate domain discrepancies.
Outcome:	The proposed framework outperforms baselines on attribute relevance and textual quality while maintaining a high inference speed.

LATENTLOGIC: Learning Logic Rules in Latent Space over Knowledge Graphs (2023.findings-emnlp)

Copied to clipboard

Junnan Liu, Qianren Mao, Chenghua Lin, Yangqiu Song, Jianxin Li

Green & Sustainable NLP

Challenge:	Existing methods for learning logic rules for knowledge graph reasoning face limitations such as searching in vast search space and inefficient optimization.
Approach:	They propose a framework to efficiently mine logic rules by controllable generation in the latent space by a pre-trained VAE and a discriminator.
Outcome:	The proposed framework efficiently mines logic rules by controllable generation in the latent space.

Effective Data Augmentation for Sentence Classification Using One VAE per Class (2022.coling-1)

Copied to clipboard

Frédéric Piedboeuf, Philippe Langlais

Learning Robust Latent Representations for Controllable Speech Synthesis (2021.findings-acl)

Challenge:	Variational auto-encoders and its conditional variant the Conditional-VAE (CVAE) are often used to generate new textual data, but they require more complex manipulations to ensure that the generated examples are useful.
Approach:	They propose a simple way to use Variational Auto-Encoders (VAE) for data augmentation by training one VAE per class.
Outcome:	The proposed method outperforms generative models on binary classification tasks and several dataset sizes on four different tasks.

Copied to clipboard

Shakti Kumar, Jithin Pradeep, Hussain Zaidi

On the Encoder-Decoder Incompatibility in Variational Text Modeling and Beyond (2020.acl-main)

Challenge:	Variational Auto-Encoders (VAEs) for learning disentangled latent representations in speech fail to learn latent clusters of speaker attributes when trained on limited or noisy datasets.
Approach:	They propose a Variational Auto-Encoder (VAE) that minimizes mutual information between latent variables and learns controllable latent representations in speech data.
Outcome:	The proposed model reduces the cluster overlap of speaker attributes by 30% over LSTM-VAE.

Copied to clipboard

Chen Wu, Prince Zizhuang Wang, William Yang Wang

Neural Topic Model with Reinforcement Learning (D19-1)

Challenge:	Existing work has shown that the optimization of variational autoencoders suffers from the posterior collapse problem.
Approach:	They propose a variational autoencoder that couples a VAE model with a deterministic autoencoding model and improves the parameters via weight sharing and decoder signal matching.
Outcome:	The proposed model improves on benchmark datasets and improves diversity of dialogue generation.

Copied to clipboard

Lin Gui, Jia Leng, Gabriele Pergola, Yu Zhou, Ruifeng Xu, Yulan He

Do sequence-to-sequence VAEs learn global features of sentences? (2020.emnlp-main)

Challenge:	Experimental results show superior performance on perplexity and topic coherence measures compared to state-of-the-art topic models.
Approach:	They propose to incorporate topic coherence measures as reward signals to guide the learning of a VAE-based topic model.
Outcome:	The proposed model is able to separating background words dynamically from topic words eliminating the pre-processing step of filtering infrequent and/or top frequent words, typically required for learning traditional topic models.

Copied to clipboard

Tom Bosc, Pascal Vincent

A Causal Approach for Counterfactual Reasoning in Narratives (2024.acl-long)

Challenge:	Autoregressive language models are often trained without explicit conditioning labels . authors question claim that latent vectors can capture global features in unsupervised manner .
Approach:	They propose to use a sequence-to-sequence architecture to learn latent variables . they find that VAEs are prone to memorizing the first words and sentence length .
Outcome:	The proposed model is prone to memorizing the first words and sentence length, the authors show . et al., 2016: a new model learns latent variables that are more global, more predictive of topic or topic labels . authors question this claim, say it is a waste of time and money .

Copied to clipboard

Feiteng Mu, Wenjie Li

Reasoning

Challenge:	Existing methods for counterfactual reasoning in narratives are based on dataset-specific heuristics, but they are abusing unique patterns, i.e., the feature of minimum editing, in the dataset, which limits the generality of their methods.
Approach:	They propose a basic VAE module for counterfactual reasoning in narratives and introduce a pre-trained classifier and external event commonsense to mitigate the posterior collapse problem.
Outcome:	The proposed method improves the causality between the counterfactual condition and the generated counterf actual outcome on two public benchmarks.

Continuous Language Generative Flow (2021.acl-long)

Copied to clipboard

Zineng Tang, Shiyue Zhang, Hyounghun Kim, Mohit Bansal

A Deep Decomposable Model for Disentangling Syntax and Semantics in Sentence Representation (2021.findings-emnlp)

Challenge:	Recent years have witnessed various types of generative models for natural language generation (NLG), especially RNNs or transformers.
Approach:	They propose a flow-based language generation model that adapts flow-derived generative models to language generation via continuous input embeddings, adapted affine coupling structures, and a novel architecture for autoregressive text generation.
Outcome:	The proposed model improves on QG and NMT and improves performance over baselines on SQuAD and TVQA and NML16.

Copied to clipboard

Dingcheng Li, Hongliang Fei, Shaogang Ren, Ping Li

A Surprisingly Effective Fix for Deep Latent Variable Modeling of Text (D19-1)

Challenge:	Recent advances in disentanglement work on coarse levels in the disenanglement of closely related properties, such as syntax and semantics in human languages.
Approach:	They propose a deep decomposable model based on VAE to disentangle syntax and semantics by using total correlation penalties on KL divergences.
Outcome:	The proposed model significantly improves the disentanglement quality between syntactic and semantic representations for semantic similarity tasks and syntaktic similarity task.

Copied to clipboard

Bohan Li, Junxian He, Graham Neubig, Taylor Berg-Kirkpatrick, Yiming Yang

Optimus: Organizing Sentences via Pre-trained Modeling of a Latent Space (2020.emnlp-main)

Challenge:	Variational Autoencoders are powerful language models and effective representation learning frameworks.
Approach:	They propose a fix for posterior collapse which improves held-out likelihood, reconstruction and latent representation learning .
Outcome:	The proposed fix significantly improves held-out likelihood, reconstruction, and latent representation learning compared with previous state-of-the-art methods.

Copied to clipboard

Chunyuan Li, Xiang Gao, Yuan Li, Baolin Peng, Xiujun Li, Yizhe Zhang, Jianfeng Gao

Implicit Deep Latent Variable Models for Text Generation (D19-1)

Challenge:	Existing models for language understanding and understanding can be trained to provide contextualized representations of words based on text data.
Approach:	They propose a large-scale language VAE model Optimus that is pre-trained on large text corpus and fine-tuned for various language generation and understanding tasks.
Outcome:	The proposed model achieves new state-of-the-art on VAE language modeling benchmarks.

Copied to clipboard

Le Fang, Chunyuan Li, Jianfeng Gao, Wen Dong, Changyou Chen

Stochastic Wasserstein Autoencoder for Probabilistic Sentence Generation (N19-1)

Challenge:	Variational auto-encoders have been used for text generation but their representation power is limited due to two reasons.
Approach:	They advocate sample-based representations of variational distributions for natural language . they further develop an LVM to directly match the aggregated posterior to the prior .
Outcome:	The proposed model can be viewed as a natural extension of VAEs with a regularization of maximizing mutual information, mitigating the "posterior collapse" issue.

Copied to clipboard

Hareesh Bahuleyan, Lili Mou, Hao Zhou, Olga Vechtomova

Enhancing Variational Autoencoders with Mutual Information Neural Estimation for Text Generation (D19-1)

Challenge:	Experimental results show that the latent space learned by WAE exhibits properties of continuity and smoothness as in VAEs.
Approach:	They propose to use the variational autoencoder (VAE) for probabilistic sentence generation . they propose a variant of WAE that encourages the stochasticity of the encoder .
Outcome:	The proposed variant encourages the stochasticity of the encoder while achieving higher BLEU scores.

Copied to clipboard

Dong Qian, William K. Cheung

A Span-based Multimodal Variational Autoencoder for Semi-supervised Multimodal Named Entity Recognition (2022.emnlp-main)

Challenge:	Existing approaches to train variational autoencoders (VAEs) have been proposed to alleviate the posterior collapse issue in NLP tasks.
Approach:	They propose to introduce a mutual information term between the input and its latent variable to regularize the objective of the VAE.
Outcome:	The proposed model performs better on three benchmark datasets and is comparable to state-of-the-art models.

Copied to clipboard

Baohang Zhou, Ying Zhang, Kehui Song, Wenya Guo, Guoqing Zhao, Hongbin Wang, Xiaojie Yuan

Multimodality

Challenge:	Existing methods for named entity recognition on social media are not efficient for semi-supervised MNER because of the mismatch between the posted text and image.
Approach:	They propose a novel method to fuse the text and image features for multimodal named entity recognition under semi-supervised setting by exploiting modal-specific VAEs.
Outcome:	The proposed method outperforms baselines under supervised setting and improves performance with less labeled data than existing semi-supervised methods.

DiffusionDialog: A Diffusion Model for Diverse Dialog Generation with Latent Space (2024.lrec-main)

Copied to clipboard

Jianxiang Xiang, Zhenhua Liu, Haodong Liu, Yin Bai, Jia Cheng, Wenliang Chen

Unsupervised Hierarchical Topic Modeling via Anchor Word Clustering and Path Guidance (2024.findings-emnlp)

Challenge:	Existing studies have tried to introduce discrete or Gaussian-based latent variables to address the one-to-many problem, but the diversity is limited.
Approach:	They propose a diffusion model to enhance the diversity of dialogue generation by using continuous latent variables instead of discrete ones.
Outcome:	The proposed model greatly enhances diversity of dialog response while keeping the coherence.

Copied to clipboard

Jiyuan Liu, Hegang Chen, Chunjiang Zhu, Yanghui Rao

Neural Gaussian Copula for Variational Autoencoder (D19-1)

Challenge:	Existing hierarchical topic models often ignore the role of anchor words that guide text generation.
Approach:	They propose to use a clustering algorithm to detect anchor words that are highly consistent with every topic and add a causal path to the popular Variational Auto-Encoder framework.
Outcome:	The proposed model outperforms state-of-the-art methods on three datasets.

Copied to clipboard

Prince Zizhuang Wang, William Yang Wang

Recurrence Boosts Diversity! Revisiting Recurrent Latent Variable in Transformer-Based Variational AutoEncoder for Diverse Text Generation (2022.findings-emnlp)

Challenge:	Variational language models assume the posterior of latent variables to be factorized even when the true posterior is not.
Approach:	They propose a Gaussian Copula Variational Autoencoder to avert a typical training problem called posterior collapse observed in all other variational language models.
Outcome:	The proposed model achieves great success over a huge number of tasks, such as transfer learning, unsupervised learning and unsupervised training.

Copied to clipboard

Jinyi Hu, Xiaoyuan Yi, Wenhao Li, Maosong Sun, Xing Xie

Causal Discovery Inspired Unsupervised Domain Adaptation for Emotion-Cause Pair Extraction (2024.findings-emnlp)

Challenge:	Variational Auto-Encoder (VAE) has been widely adopted in text generation due to its ability to learn flexible representations.
Approach:	They propose a Transformer-based recurrent VAE structure that imposes recurrence on segment-wise latent variables with arbitrarily separated text segments and constructs the posterior distribution with residual parameterization.
Outcome:	The proposed structure can deduce a non-zero lower bound of the KL term and enhance the entanglement of each segment and preceding latent variables, providing a theoretical guarantee of generation diversity.

Copied to clipboard

Yuncheng Hua, Yujin Huang, Shuo Huang, Tao Feng, Lizhen Qu, Christopher Bain, Richard Bassed, Reza Haf

Low-Resource NLP

Challenge:	Emotion-cause pair extraction is a task that aims to extract emotions and the events causing such emotions.
Approach:	They propose a deep latent model which captures the underlying latent structures of data and utilizes the easily transferable knowledge of emotions as the bridge to link the distributions of events in different domains.
Outcome:	The proposed model outperforms the strongest baseline by approximately 11.05% on a Chinese benchmark and 2.45% on an English benchmark in terms of weighted-average F1 score.

VaeDiff-DocRE: End-to-end Data Augmentation Framework for Document-level Relation Extraction (2025.coling-main)

Copied to clipboard

Khai Phan Tran, Wen Hua, Xue Li

Coherent Entity Disambiguation via Modeling Topic and Categorical Dependency (2023.findings-emnlp)

Challenge:	Existing methods for Document-level Relation Extraction assume a uniform label distribution, resulting in suboptimal performance on real-world, imbalanced datasets.
Approach:	They propose a method that leverages the Variational Autoencoder architecture to capture all relation-wise distributions formed by entity pair representations and augment data for underrepresented relations.
Outcome:	The proposed method outperforms state-of-the-art models on two benchmark datasets and is available on github.

Copied to clipboard

Zilin Xiao, Linjun Shou, Xingyao Zhang, Jie Wu, Ming Gong, Daxin Jiang

Document Hashing with Mixture-Prior Generative Models (D19-1)

Challenge:	Existing entity disambiguation methods struggle to capture explicit discourse-level dependencies, resulting in incoherent predictions at the abstract level.
Approach:	They propose an unsupervised variational autoencoder to extract latent topic vectors of context sentences to enhance coherence of entity predictions.
Outcome:	The proposed system achieves state-of-the-art on popular ED benchmarks with an average improvement of 1.3 F1 points.

Copied to clipboard

Wei Dong, Qinliang Su, Dinghan Shen, Changyou Chen

Challenge:	Existing generative hashing methods only consider the use of simple priors, which limits them to further improve their performance.
Approach:	They propose to use Gaussian and Bernoulli priors to generate hashing codes . they propose to cast a Gausssian latent representation into binary code .
Outcome:	The proposed models outperform existing methods on a benchmark dataset using Gaussian and Bernoulli priors.

Contrastive Deterministic Autoencoders For Language Modeling (2023.findings-emnlp)

Copied to clipboard

Amur Ghose, Pascal Poupart

Decode with Template: Content Preserving Sentiment Transfer (2020.lrec-1)

Challenge:	Variational autoencoders (VAEs) are a popular family of generative models with wide applicability.
Approach:	They propose to modify a deterministic model designed for images to avoid posterior collapse by controlling the entropy of the aggregate posterior to make it Gaussian.
Outcome:	The proposed models outperform a broad range of VAE models on text generation and downstream tasks from representations while avoiding reparametrization steps.

Copied to clipboard

Zhiyuan Wen, Jiannong Cao, Ruosong Yang, Senzhang Wang

Sentiment Analysis

Challenge:	Existing methods to transfer sentiments for text use only explicit sentiments and templates to remove them from input sentences.
Approach:	They propose a method to transfer sentiments from input sentences to output sentences using templates.
Outcome:	The proposed model significantly outperforms state-of-the-art models in content preservation.

Generating Sentences from Disentangled Syntactic and Semantic Spaces (P19-1)

Copied to clipboard

Yu Bao, Hao Zhou, Shujian Huang, Lei Li, Lili Mou, Olga Vechtomova, Xin-yu Dai, Jiajun Chen

Syntactic Text Processing

Challenge:	Variational auto-encoders (VAEs) are widely used in natural language generation due to the regularization of the latent space.
Approach:	They propose to generate sentences from disentangled syntactic and semantic spaces by using the linearized tree sequence.
Outcome:	The proposed method achieves similar or better performance in various tasks compared with state-of-the-art models.

Variational Autoencoder with Disentanglement Priors for Low-Resource Task-Specific Natural Language Generation (2022.emnlp-main)

Copied to clipboard

Zhuang Li, Lizhen Qu, Qiongkai Xu, Tongtong Wu, Tianyang Zhan, Gholamreza Haffari

Low-Resource NLP

Challenge:	Existing models for task-specific natural language generation do not contain any labeled examples.
Approach:	They propose a variational autoencoder with disentanglement priors for task-specific natural language generation with none or a handful of task-related labeled examples.
Outcome:	The proposed model outperforms baseline models in terms of data augmentation and text style transfer in the few-shot setting.

Prompt Conditioned VAE: Enhancing Generative Replay for Lifelong Learning in Task-Oriented Dialogue (2022.emnlp-main)

Copied to clipboard

Yingxiu Zhao, Yinhe Zheng, Zhiliang Tian, Chang Gao, Jian Sun, Nevin L. Zhang

Empirical Prior for Text Autoencoders (2024.findings-emnlp)

Challenge:	Existing generative replay methods use only a single task-specific token to control their models.
Approach:	They propose a method to capture task-specific distributions with a conditional variational autoencoder, conditioned on natural language prompts to guide the pseudo-sample generation.
Outcome:	The proposed method outperforms baselines on natural language understanding tasks of advanced task-oriented dialogue (ToD) systems.

Copied to clipboard

Yongjing Yin, Wenyang Gao, Haodong Wu, Jianhao Yan, Yue Zhang

Combining Hierachical VAEs with LLMs for clinically meaningful timeline summarisation in social media (2024.findings-acl)

Challenge:	Variational Autoencoders (VAE) are used to train generative models with latent variables.
Approach:	They propose a transition from Variational Autoencoders (VAE) to text autoencodeurs (AE) which model a compact latent space and preserves the capability of the language model itself.
Outcome:	The proposed method generates higher quality and more diverse text than the VAE-based Transformer baselines, and is more efficient than previous approaches.

Copied to clipboard

Jiayu Song, Jenny Chim, Adam Tsakalidis, Julia Ive, Dana Atzil-Slonim, Maria Liakata

Tokenization with Factorized Subword Encoding (2023.findings-acl)

Challenge:	Existing studies have shown that social media users' posts can help identify depression, bipolar disorder or self-harm.
Approach:	They propose a hybrid abstractive summarisation approach combining hierarchical VAEs with LLMs to produce clinically meaningful summaries from social media timelines.
Outcome:	The proposed approach produces clinically meaningful summaries from social media user timelines, suitable for mental health monitoring.

Copied to clipboard

David Samuel, Lilja Øvrelid

Syntactic Text Processing

Challenge:	Subword tokenization methods are often used to project subwords onto triplets . a typical tokenizer consists of 10 000s of subword mapped onto a single index .
Approach:	They propose a subword tokenization method that factorizes subwords onto triplets using a VQ-VAE model.
Outcome:	The proposed tokenization method is more appropriate and robust for morphological tasks than the commonly used byte-pair encoding (BPE) tokenization algorithm.

DiffLM: Controllable Synthetic Data Generation via Diffusion Language Models (2025.findings-acl)

Copied to clipboard

Ying Zhou, Xinyao Wang, Yulei Niu, Yaojie Shen, Lexin Tang, Fan Chen, Ben He, Le Sun, Longyin Wen

S2WTM: Spherical Sliced-Wasserstein Autoencoder for Topic Modeling (2025.acl-long)

Challenge:	Recent advances in large language models (LLMs) have significantly enhanced their knowledge and generative capabilities, leading to a surge of interest in leveraging LLMs for high-quality data synthesis.
Approach:	They propose a controllable data synthesis framework based on variational autoencoder which leverages diffusion models to reserve more information of original distribution and format structure in the learned latent distribution.
Outcome:	The proposed framework generates high-quality data with performance exceeding that of real data by 2%–7% on seven real-world datasets.

Copied to clipboard

Suman Adhya, Debarshi Kumar Sanyal

Capturing Latent Modal Association For Multimodal Entity Alignment (2025.findings-emnlp)

Challenge:	Variational autoencoder-based neural topic models lack posterior collapse in high-dimensional text data.
Approach:	They propose a Spherical Sliced Wasserstein Autoencoder for Topic Modeling to model latent representations in a hyperspherical space.
Outcome:	The proposed model outperforms state-of-the-art topic models while improving performance on downstream tasks.

Copied to clipboard

Yongquan Ji, Jingwei Cheng, Fu Zhang, Chenglong Lu

Scale-VAE: Preventing Posterior Collapse in Variational Autoencoder (2024.lrec-main)

Challenge:	Existing methods for multimodal entity alignment overlook the quality of input modality embeddings during modality interaction, amplifying noise propagation while suppressing discriminative feature representations.
Approach:	They propose a model for capturing latent modal association for multimodal entity alignment using a self-attention mechanism to enhance salient information while attenuating noise within individual modality embeddings.
Outcome:	The proposed model achieves an absolute 3.1% higher Hits@1 score than the sota method.

Copied to clipboard

Tianbao Song, Jingbo Sun, Xin Liu, Weiming Peng

Reusable Experiences: Latent Routing and Modular Composition in LLMs (2026.acl-long)

Challenge:	Variational autoencoder (VAE) is a widely used generative model . but when employing strong autoregressive generation networks, VAE tends to converge to a degenerate local optimum known as posterior collapse.
Approach:	They propose a model called Scale-VAE to solve a posterior collapse problem . they use a factor to keep the posterior dimension discriminative across data instances .
Outcome:	The proposed model outperforms state-of-the-art models in density estimation and representation learning.

Copied to clipboard

Shuai Ling, Lizi Liao, Dongmei Jiang, Weili Guan

V-VAE: A Variational Auto Encoding Framework Towards Fine-Grained Control over Human-Like Chat (2025.emnlp-main)

Challenge:	Existing approaches represent accumulated experience as explicit textual artifacts in prompts or implicitly within model weights via fine-tuning. Existing methods are limited by context windows and cannot internalize knowledge.
Approach:	They propose a framework that treats latent experiences as fundamental units for LLM specialization.
Outcome:	Experiments on multi-task NLP benchmarks show that this approach outperforms standard fine-tuning, yielding improved generalization through flexible skill reuse.

Copied to clipboard

Qi Lin, Weikai Xu, Lisi Chen, Bin Dai