Masked Language Model Research Articles

Unifying Multi-Modal Uncertainty Modeling and Semantic Alignment for Text-to-Image Person Re-identification

Text-to-Image person re-identification (TI-ReID) aims to retrieve the images of target identity according to the given textual description. The existing methods in TI-ReID focus on aligning the visual and textual modalities through contrastive feature alignment or reconstructive masked language modeling (MLM). However, these methods parameterize the image/text instances as deterministic embeddings and do not explicitly consider the inherent uncertainty in pedestrian images and their textual descriptions, leading to limited image-text relationship expression and semantic alignment. To address the above problem, in this paper, we propose a novel method that unifies multi-modal uncertainty modeling and semantic alignment for TI-ReID. Specifically, we model the image and textual feature vectors of pedestrian as Gaussian distributions, where the multi-granularity uncertainty of the distribution is estimated by incorporating batch-level and identity-level feature variances for each modality. The multi-modal uncertainty modeling acts as a feature augmentation and provides richer image-text semantic relationship. Then we present a bi-directional cross-modal circle loss to more effectively align the probabilistic features between image and text in a self-paced manner. To further promote more comprehensive image-text semantic alignment, we design a task that complements the masked language modeling, focusing on the cross-modality semantic recovery of global masked token after cross-modal interaction. Extensive experiments conducted on three TI-ReID datasets highlight the effectiveness and superiority of our method over state-of-the-arts.

Robust Evaluation Measures for Evaluating Social Biases in Masked Language Models

Many evaluation measures are used to evaluate social biases in masked language models (MLMs). However, we find that these previously proposed evaluation measures are lacking robustness in scenarios with limited datasets. This is because these measures are obtained by comparing the pseudo-log-likelihood (PLL) scores of the stereotypical and anti-stereotypical samples using an indicator function. The disadvantage is the limited mining of the PLL score sets without capturing its distributional information. In this paper, we represent a PLL score set as a Gaussian distribution and use Kullback-Leibler (KL) divergence and Jensen–Shannon (JS) divergence to construct evaluation measures for the distributions of stereotypical and anti-stereotypical PLL scores. Experimental results on the publicly available datasets StereoSet (SS) and CrowS-Pairs (CP) show that our proposed measures are significantly more robust and interpretable than those proposed previously.

A Multimodal, Multi-Task Adapting Framework for Video Action Recognition

Recently, the rise of large-scale vision-language pretrained models like CLIP, coupled with the technology of Parameter-Efficient FineTuning (PEFT), has captured substantial attraction in video action recognition. Nevertheless, prevailing approaches tend to prioritize strong supervised performance at the expense of compromising the models' generalization capabilities during transfer. In this paper, we introduce a novel Multimodal, Multi-task CLIP adapting framework named M2-CLIP to address these challenges, preserving both high supervised performance and robust transferability. Firstly, to enhance the individual modality architectures, we introduce multimodal adapters to both the visual and text branches. Specifically, we design a novel visual TED-Adapter, that performs global Temporal Enhancement and local temporal Difference modeling to improve the temporal representation capabilities of the visual encoder. Moreover, we adopt text encoder adapters to strengthen the learning of semantic label information. Secondly, we design a multi-task decoder with a rich set of supervisory signals, including the original contrastive learning head, a cross-modal classification head, a cross-modal masked language modeling head, and a visual classification head. This multi-task decoder adeptly satisfies the need for strong supervised performance within a multimodal framework. Experimental results validate the efficacy of our approach, demonstrating exceptional performance in supervised learning while maintaining strong generalization in zero-shot scenarios.

NanoBERTa-ASP: predicting nanobody paratope based on a pretrained RoBERTa model

BackgroundNanobodies, also known as VHH or single-domain antibodies, are unique antibody fragments derived solely from heavy chains. They offer advantages of small molecules and conventional antibodies, making them promising therapeutics. The paratope is the specific region on an antibody that binds to an antigen. Paratope prediction involves the identification and characterization of the antigen-binding site on an antibody. This process is crucial for understanding the specificity and affinity of antibody-antigen interactions. Various computational methods and experimental approaches have been developed to predict and analyze paratopes, contributing to advancements in antibody engineering, drug development, and immunotherapy. However, existing predictive models trained on traditional antibodies may not be suitable for nanobodies. Additionally, the limited availability of nanobody datasets poses challenges in constructing accurate models.MethodsTo address these challenges, we have developed a novel nanobody prediction model, named NanoBERTa-ASP (Antibody Specificity Prediction), which is specifically designed for predicting nanobody-antigen binding sites. The model adopts a training strategy more suitable for nanobodies, based on an advanced natural language processing (NLP) model called BERT (Bidirectional Encoder Representations from Transformers). To be more specific, the model utilizes a masked language modeling approach named RoBERTa (Robustly Optimized BERT Pretraining Approach) to learn the contextual information of the nanobody sequence and predict its binding site.ResultsNanoBERTa-ASP achieved exceptional performance in predicting nanobody binding sites, outperforming existing methods, indicating its proficiency in capturing sequence information specific to nanobodies and accurately identifying their binding sites. Furthermore, NanoBERTa-ASP provides insights into the interaction mechanisms between nanobodies and antigens, contributing to a better understanding of nanobodies and facilitating the design and development of nanobodies with therapeutic potential.ConclusionNanoBERTa-ASP represents a significant advancement in nanobody paratope prediction. Its superior performance highlights the potential of deep learning approaches in nanobody research. By leveraging the increasing volume of nanobody data, NanoBERTa-ASP can further refine its predictions, enhance its performance, and contribute to the development of novel nanobody-based therapeutics.Github repository: https://github.com/WangLabforComputationalBiology/NanoBERTa-ASP

Bidirectional encoders to state-of-the-art: a review of BERT and its transformative impact on natural language processing

First developed in 2018 by Google researchers, Bidirectional Encoder Representations from Transformers (BERT) represents a breakthrough in natural language processing (NLP). BERT achieved state-of-the-art results across a range of NLP tasks while using a single transformer-based neural network architecture. This work reviews BERT's technical approach, performance when published, and significant research impact since release. We provide background on BERT's foundations like transformer encoders and transfer learning from universal language models. Core technical innovations include deeply bidirectional conditioning and a masked language modeling objective during BERT's unsupervised pretraining phase. For evaluation, BERT was fine-tuned and tested on eleven NLP tasks ranging from question answering to sentiment analysis via the GLUE benchmark, achieving new state-of-the-art results. Additionally, this work analyzes BERT's immense research influence as an accessible technique surpassing specialized models. BERT catalyzed adoption of pretraining and transfer learning for NLP. Quantitatively, over 10,000 papers have extended BERT and it is integrated widely across industry applications. Future directions based on BERT scale towards billions of parameters and multilingual representations. In summary, this work reviews the method, performance, impact and future outlook for BERT as a foundational NLP technique. We provide background on BERT's foundations like transformer encoders and transfer learning from universal language models. Core technical innovations include deeply bidirectional conditioning and a masked language modeling objective during BERT's unsupervised pretraining phase. For evaluation, BERT was fine-tuned and tested on eleven NLP tasks ranging from question answering to sentiment analysis via the GLUE benchmark, achieving new state-of-the-art results. Additionally, this work analyzes BERT's immense research influence as an accessible technique surpassing specialized models. BERT catalyzed adoption of pretraining and transfer learning for NLP. Quantitatively, over 10,000 papers have extended BERT and it is integrated widely across industry applications. Future directions based on BERT scale towards billions of parameters and multilingual representations. In summary, this work reviews the method, performance, impact and future outlook for BERT as a foundational NLP technique.

Rethinking the Exploitation of Monolingual Data for Low-Resource Neural Machine Translation

Abstract The utilization of monolingual data has been shown to be a promising strategy for addressing low-resource machine translation problems. Previous studies have demonstrated the effectiveness of techniques such as back-translation and self-supervised objectives, including masked language modeling, causal language modeling, and denoise autoencoding, in improving the performance of machine translation models. However, the manner in which these methods contribute to the success of machine translation tasks and how they can be effectively combined remains an under-researched area. In this study, we carry out a systematic investigation of the effects of these techniques on linguistic properties through the use of probing tasks, including source language comprehension, bilingual word alignment, and translation fluency. We further evaluate the impact of pre-training, back-translation, and multi-task learning on bitexts of varying sizes. Our findings inform the design of more effective pipelines for leveraging monolingual data in extremely low-resource and low-resource machine translation tasks. Experiment results show consistent performance gains in seven translation directions, which provide further support for our conclusions and understanding of the role of monolingual data in machine translation.

Convolutions are competitive with transformers for protein sequence pretraining

Pretrained protein sequence language models have been shown to improve the performance of many prediction tasks and are now routinely integrated into bioinformatics tools. However, these models largely rely on the transformer architecture, which scales quadratically with sequence length in both run-time and memory. Therefore, state-of-the-art models have limitations on sequence length. To address this limitation, we investigated whether convolutional neural network (CNN) architectures, which scale linearly with sequence length, could be as effective as transformers in protein language models. With masked language model pretraining, CNNs are competitive with, and occasionally superior to, transformers across downstream applications while maintaining strong performance on sequences longer than those allowed in the current state-of-the-art transformer models. Our work suggests that computational efficiency can be improved without sacrificing performance, simply by using a CNN architecture instead of a transformer, and emphasizes the importance of disentangling pretraining task and model architecture. A record of this paper's transparent peer review process is included in the supplemental information.

A reversible natural language watermarking for sensitive information protection

Existing methods have evolved from using synonym substitution to incorporating arbitrary word substitution to achieve reversible natural language watermarking. However, a notable limitation is that they are prone to overlook the sensitivity of information associated with the original words, with a tendency to prefer non-sensitive words for substitution. As a result, a potential risk of sensitive information leakage contained in the original text is posed. Furthermore, while aiming for reversibility, the overall performance of the watermarking method may be inadvertently compromised. In response to the above problems, this paper puts forward a novel reversible natural language watermarking method that combines a Keyword Substitution scheme and a Prediction Error Expansion algorithm (KSPEE) to protect sensitive information, verify content integrity, protect copyright, and so on. Specifically, KSPEE leverages a keyword extraction algorithm to identify important content containing sensitive information in the original text, thereby determining the potential positions for watermark information embedding. Subsequently, a masked language model is utilized to predict appropriate substitution words based on the surrounding semantic information of the embedding position. In addition, the prediction error expansion algorithm is employed to select appropriate words for substituting the original keywords, ensuring the successful embedding of watermark information while maintaining the recoverability of the original keywords. By identifying keywords and substituting them, a suitable method of protecting the original sensitive information is provided. Extensive experiments demonstrate that, under the promise of semantic distortion and lossless restoration of the original content, the proposed method KSPEE achieves outstanding watermarked text quality. A higher watermark embedding rate is achieved and strong security is shown by KSPEE. More importantly, KSPEE effectively prevents the leakage of sensitive information.

SRBerta—A Transformer Language Model for Serbian Cyrillic Legal Texts

Language is a unique ability of human beings. Although relatively simple for humans, the ability to understand human language is a highly complex task for machines. For a machine to learn a particular language, it must understand not only the words and rules used in a particular language, but also the context of sentences and the meaning that words take on in a particular context. In the experimental development we present in this paper, the goal was the development of the language model SRBerta—a language model designed to understand the formal language of Serbian legal documents. SRBerta is the first of its kind since it has been trained using Cyrillic legal texts contained within a dataset created specifically for this purpose. The main goal of SRBerta network development was to understand the formal language of Serbian legislation. The training process was carried out using minimal resources (single NVIDIA Quadro RTX 5000 GPU) and performed in two phases—base model training and fine-tuning. We will present the structure of the model, the structure of the training datasets, the training process, and the evaluation results. Further, we will explain the accuracy metric used in our case and demonstrate that SRBerta achieves a high level of accuracy for the task of masked language modeling in Serbian Cyrillic legal texts. Finally, SRBerta model and training datasets are publicly available for scientific and commercial purposes.

A Character Based Steganography Using Masked Language Modeling

A Character Based Steganography Using Masked Language Modeling

Boosting Point-BERT by Multi-Choice Tokens

Masked language modeling (MLM) has become one of the most successful self-supervised pre-training task. Inspired by its success, Point-BERT, as a pioneer work in point cloud, proposed masked point modeling (MPM) to pre-train point transformer on large scale unanotated dataset. Despite its great performance, we find the inherent difference between language and point cloud tends to cause ambiguous tokenization for point cloud, and no gold standard is available for point cloud tokenization. Point-BERT uses a discrete Variational AutoEncoder (dVAE) as tokenizer, but it might generate different token ids for semantically-similar patches and the same token ids for semantically-dissimilar patches. To tackle the above problems, we propose our McP-BERT, a pre-training framework with multi-choice tokens. Specifically, we ease the previous single-choice constraint on patch token ids in Point-BERT, and provide multi-choice token ids for each patch as supervision. Moreover, we utilitze the high-level semantics learned by transformer to further refine our supervision signals. Extensive experiments on point cloud classification, few-shot classification and part segmentation tasks demonstrate the superiority of our method, e.g., the pre-trained transformer achieves 94.1% accuracy on ModelNet40, 84.28% accuracy on the hardest setting of ScanObjectNN and new state-of-the-art performance on few-shot learning. Our method improves the performance of Point-BERT on all down-stream tasks without extra computational overhead.

Transformer-Based High-Fidelity Facial Displacement Completion for Detailed 3D Face Reconstruction

In this paper, we tackle a special face completion task, facial displacement completion, which can offer a key component for many single image 3D face reconstruction systems. To produce a detailed 3D face with ear-to-ear complete displacement UV map, we propose a novel Displacement Completion method based on Transformer (DCT). Current transformer based image inpainting methods usually follow a two-stage scheme, which firstly recovers the masked pixels in low resolution with transformer, and then replenishes the inpainting result in high resolution with GAN. Although these methods have achieved great success, they suffer from information loss from two aspects when applied in face completion: 1) The downsampling operation makes transformer only produce a coarse appearance prior for GAN, incurring middle and low level information loss. 2) Some meaningful facial semantics can be well captured by transformer and further benefit the completion, but it's has not yet been explored. Motivated by the above considerations, we come up with three key designs in the proposed DCT: PCA tokenization, BERT-style learning, and style modulation. Firstly, we use PCA tokenization to replace the downsampling in transformer to preserve more meaningful structures. Secondly, we make transformer simulate the two tasks in BERT, Masked Language Model (MLM) and Next Sentence Prediction (NSP), for both masked pixels and facial attributes recovery. Thirdly, we encode the outcome of transformer as the latent code to guide an image translation network in the StyleGAN2 modulation way. Experments on both FaceScape dataset and in-the-wild data demonstrate DCT's better performance compared with other transformer based or GAN based completion methods.

Efficient Masked Autoencoders With Self-Consistency.

Inspired by the masked language modeling (MLM) in natural language processing tasks, the masked image modeling (MIM) has been recognized as a strong self-supervised pre-training method in computer vision. However, the high random mask ratio of MIM results in two serious problems: 1) the inadequate data utilization of images within each iteration brings prolonged pre-training, and 2) the high inconsistency of predictions results in unreliable generations, i.e., the prediction of the identical patch may be inconsistent in different mask rounds, leading to divergent semantics in the ultimately generated outcomes. To tackle these problems, we propose the efficient masked autoencoders with self-consistency (EMAE) to improve the pre-training efficiency and increase the consistency of MIM. In particular, we present a parallel mask strategy that divides the image into K non-overlapping parts, each of which is generated by a random mask with the same mask ratio. Then the MIM task is conducted parallelly on all parts in an iteration and the model minimizes the loss between the predictions and the masked patches. Besides, we design the self-consistency learning to further maintain the consistency of predictions of overlapping masked patches among parts. Overall, our method is able to exploit the data more efficiently and obtains reliable representations. Experiments on ImageNet show that EMAE achieves the best performance on ViT-Large with only 13% of MAE pre-training time using NVIDIA A100 GPUs. After pre-training on diverse datasets, EMAE consistently obtains state-of-the-art transfer ability on a variety of downstream tasks, such as image classification, object detection, and semantic segmentation.

Personalized Image Enhancement Featuring Masked Style Modeling

We address personalized image enhancement in this study, where we enhance input images for each user based on the user's preferred images. Previous methods apply the same preferred style to all input images (i.e., only one style for each user); in contrast to these methods, we aim to achieve content-aware personalization by applying different styles to each image considering the contents. For content-aware personalization, we make two contributions. First, we propose a method named masked style modeling, which can predict a style for an input image considering the contents by using the framework of masked language modeling. Second, to allow this model to consider the contents of images, we propose a novel training scheme where we download images from Flickr and create pseudo input and retouched image pairs using a degrading model. We conduct quantitative evaluations and a user study, and our method trained using our training scheme successfully achieves content-aware personalization; moreover, our method outperforms other previous methods in this field. Our source code is available at https://github.com/satoshi-kosugi/masked-style-modeling.

Two-Step Masked Language Model for Domain-Adapting Multi-Modal Task-Oriented Dialogue Systems

Two-Step Masked Language Model for Domain-Adapting Multi-Modal Task-Oriented Dialogue Systems

MvSR-NAT: Multi-view Subset Regularization for Non-Autoregressive Machine Translation

Conditional masked language models (CMLM) have shown impressive progress in non-autoregressive machine translation (NAT). They learn the conditional translation model by predicting the random masked subset in the target sentence. Based on the CMLM framework, we introduce Multi-view Subset Regularization (MvSR), a novel regularization method to improve the performance of the NAT model. Specifically, MvSR consists of two parts: (1) <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">shared mask consistency</i> : we forward the same target with different mask strategies, and encourage the predictions of shared mask positions to be consistent with each other. (2) <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">model consistency</i> , we maintain an exponential moving average of the model weights, and enforce the predictions to be consistent between the average model and the online model. Without changing the CMLM-based architecture, our approach achieves remarkable performance on three public benchmarks with 0.7-1.15 BLEU gains over previous NAT models. And, we reduce the gap to the stronger Transformer baseline.

Efficient Text Style Transfer through Robust Masked Language Model and Iterative Inference

Efficient Text Style Transfer through Robust Masked Language Model and Iterative Inference

Enhancing Generic Reaction Yield Prediction through Reaction Condition-Based Contrastive Learning.

Deep learning (DL)-driven efficient synthesis planning may profoundly transform the paradigm for designing novel pharmaceuticals and materials. However, the progress of many DL-assisted synthesis planning (DASP) algorithms has suffered from the lack of reliable automated pathway evaluation tools. As a critical metric for evaluating chemical reactions, accurate prediction of reaction yields helps improve the practicality of DASP algorithms in the real-world scenarios. Currently, accurately predicting yields of interesting reactions still faces numerous challenges, mainly including the absence of high-quality generic reaction yield datasets and robust generic yield predictors. To compensate for the limitations of high-throughput yield datasets, we curated a generic reaction yield dataset containing 12 reaction categories and rich reaction condition information. Subsequently, by utilizing 2 pretraining tasks based on chemical reaction masked language modeling and contrastive learning, we proposed a powerful bidirectional encoder representations from transformers (BERT)-based reaction yield predictor named Egret. It achieved comparable or even superior performance to the best previous models on 4 benchmark datasets and established state-of-the-art performance on the newly curated dataset. We found that reaction-condition-based contrastive learning enhances the model's sensitivity to reaction conditions, and Egret is capable of capturing subtle differences between reactions involving identical reactants and products but different reaction conditions. Furthermore, we proposed a new scoring function that incorporated Egret into the evaluation of multistep synthesis routes. Test results showed that yield-incorporated scoring facilitated the prioritization of literature-supported high-yield reaction pathways for target molecules. In addition, through meta-learning strategy, we further improved the reliability of the model's prediction for reaction types with limited data and lower data quality. Our results suggest that Egret holds the potential to become an essential component of the next-generation DASP tools.

POS-BERT: Point cloud one-stage BERT pre-training

Recently, the pre-training paradigm combining Transformer and masked language modeling in BERT has achieved tremendous success not only in NLP, but also in images and point clouds. However, directly extending BERT from NLP to point clouds requires first training a discrete Variational AutoEncoder (dVAE) as the tokenizer, which results in a complex two-stage process, as in Point-BERT. Inspired by BERT and MoCo, we propose POS-BERT, a one-stage BERT pre-training method for point clouds. Specifically, we use the masked patch modeling (MPM) task to perform point cloud pre-training, which aims to recover masked patch information under the supervision of a tokenizer’s output. Unlike Point-BERT, whose tokenizer is extra-trained and frozen, we propose a momentum tokenizer which is dynamically updated during training the Transformer. Furthermore, in order to better learn high-level semantic representation, we integrate contrastive learning into the proposed framework to maximize the class token consistency between augmented point cloud pairs. Experiments show that POS-BERT achieves the state-of-the-art performance on linear SVM classification of ModelNet40 with fixed feature extractors, and it exceeds Point-BERT by 3.5%. In addition, POS-BERT has significantly improved many downstream tasks, including fine-tuned classification, few-shot classification and part segmentation. The code and trained models will be released on https://github.com/fukexue/POS-BERT.git.

Deep learning workflow for the inverse design of molecules with specific optoelectronic properties

The inverse design of novel molecules with a desirable optoelectronic property requires consideration of the vast chemical spaces associated with varying chemical composition and molecular size. First principles-based property predictions have become increasingly helpful for assisting the selection of promising candidate chemical species for subsequent experimental validation. However, a brute-force computational screening of the entire chemical space is decidedly impossible. To alleviate the computational burden and accelerate rational molecular design, we here present an iterative deep learning workflow that combines (i) the density-functional tight-binding method for dynamic generation of property training data, (ii) a graph convolutional neural network surrogate model for rapid and reliable predictions of chemical and physical properties, and (iii) a masked language model. As proof of principle, we employ our workflow in the iterative generation of novel molecules with a target energy gap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO).