More Like this

1. What Can an Accent Identifier Learn? Probing Phonetic and Prosodic Information in a Wav2vec2-based Accent Identification Model

   https://doi.org/10.21437/Interspeech.2023-2254  

   Yang, Mu; Shekar, Ram C.; Kang, Okim; Hansen, John H. (August 2023, ISCA INTERSPEECH-2023)

   N/A (Ed.)

   This study is focused on understanding and quantifying the change in phoneme and prosody information encoded in the Self-Supervised Learning (SSL) model, brought by an accent identification (AID) fine-tuning task. This problem is addressed based on model probing. Specifically, we conduct a systematic layer-wise analysis of the representations of the Transformer layers on a phoneme correlation task, and a novel word-level prosody prediction task. We compare the probing performance of the pre-trained and fine-tuned SSL models. Results show that the AID fine-tuning task steers the top 2 layers to learn richer phoneme and prosody representation. These changes share some similarities with the effects of fine-tuning with an Automatic Speech Recognition task. In addition, we observe strong accent-specific phoneme representations in layer 9. To sum up, this study provides insights into the understanding of SSL features and their interactions with fine-tuning tasks. 

   more » « less
2. Topology-Informed Pre-training of Graph Neural Networks

   https://doi.org/10.1109/ICASSP49660.2025.10890271  

   Liang, Peiyu; Gel, Yulia R; Chen, Yuzhou (April 2025, IEEE)

   Pre-training has emerged as a dominant paradigm in graph representation learning to address data scarcity and generalization challenges. The majority of existing methods primarily focus on refining fine-tuning and prompting techniques to extract information from pre-trained models. However, the effectiveness of these approaches is contingent upon the quality of the pre-trained knowledge (i.e., latent representations). Inspired by the recent success in topological representation learning, we propose a novel pre-training strategy to capture and learn topological information of graphs. The key to the success of our strategy is to pre-train expressive Graph Neural Networks (GNNs) at the levels of individual nodes while accounting for the key topological characteristics of a graph so that GNNs become sufficiently powerful to effectively encode input graph information. The proposed model is designed to be seamlessly integrated with various downstream graph representation learning tasks. 

   more » « less
3. Integrating Multimodal Information in Large Pretrained Transformers

   Rahman, Wasifur; Hasan, Md Kamrul; Lee, Sangwu; Zadeh, AmirAli Bagher; Mao, Chengfeng; Morency, Louis-Philippe; Hoque, Ehsan (January 2020, Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics)

   Recent Transformer-based contextual word representations, including BERT and XLNet, have shown state-of-the-art performance in multiple disciplines within NLP. Fine-tuning the trained contextual models on task-specific datasets has been the key to achieving superior performance downstream. While fine-tuning these pre-trained models is straightforward for lexical applications (applications with only language modality), it is not trivial for multimodal language (a growing area in NLP focused on modeling face-to-face communication). More specifically, this is due to the fact that pre-trained models don’t have the necessary components to accept two extra modalities of vision and acoustic. In this paper, we proposed an attachment to BERT and XLNet called Multimodal Adaptation Gate (MAG). MAG allows BERT and XLNet to accept multimodal nonverbal data during fine-tuning. It does so by generating a shift to internal representation of BERT and XLNet; a shift that is conditioned on the visual and acoustic modalities. In our experiments, we study the commonly used CMU-MOSI and CMU-MOSEI datasets for multimodal sentiment analysis. Fine-tuning MAG-BERT and MAG-XLNet significantly boosts the sentiment analysis performance over previous baselines as well as language-only fine-tuning of BERT and XLNet. On the CMU-MOSI dataset, MAG-XLNet achieves human-level multimodal sentiment analysis performance for the first time in the NLP community. 

   more » « less
4. Leveraging Multiple Sources in Automatic African American English Dialect Detection for Adults and Children

   https://doi.org/10.1109/ICASSP49357.2023.10096614  

   Johnson, Alexander; Shetty, Vishwas; Ostendorf, Mari; and Alwan, Abeer (June 2023, ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP))

   IEEE SIGNAL PROCESSING SOCIETY (Ed.)

   This paper 1 presents a novel system which utilizes acoustic, phonological, morphosyntactic, and prosodic information for binary automatic dialect detection of African American English. We train this system utilizing adult speech data and then evaluate on both children’s and adults’ speech with unmatched training and testing scenarios. The proposed system combines novel and state-of-the-art architectures, including a multi-source transformer language model pre-trained on Twitter text data and fine-tuned on ASR transcripts as well as an LSTM acoustic model trained on self-supervised learning representations, in order to learn a comprehensive view of dialect. We show robust, explainable performance across recording conditions for different features for adult speech, but fusing multiple features is important for good results on children’s speech. 

   more » « less
5. Edge Prompt Tuning for Graph Neural Networks

   Fu, Xingbo; He, Yinhan; Li, Jundong (April 2025, International Conference on Learning Representations)

   Pre-training powerful Graph Neural Networks (GNNs) with unlabeled graph data in a self-supervised manner has emerged as a prominent technique in recent years. However, inevitable objective gaps often exist between pre-training and downstream tasks. To bridge this gap, graph prompt tuning techniques design and learn graph prompts by manipulating input graphs or reframing downstream tasks as pre-training tasks without fine-tuning the pre-trained GNN models. While recent graph prompt tuning methods have proven effective in adapting pre-trained GNN models for downstream tasks, they overlook the crucial role of edges in graph prompt design, which can significantly affect the quality of graph representations for downstream tasks. In this study, we propose EdgePrompt, a simple yet effective graph prompt tuning method from the perspective of edges. Unlike previous studies that design prompt vectors on node features, EdgePrompt manipulates input graphs by learning additional prompt vectors for edges and incorporates the edge prompts through message passing in the pre-trained GNN models to better embed graph structural information for downstream tasks. Our method is compatible with prevalent GNN architectures pre-trained under various pre-training strategies and is universal for different downstream tasks. We provide comprehensive theoretical analyses of our method regarding its capability of handling node classification and graph classification as downstream tasks. Extensive experiments on ten graph datasets under four pre-training strategies demonstrate the superiority of our proposed method against six baselines. 

   more » « less