An official website of the United States government
The automation of extracting argument structures faces a pair of challenges on (1) encoding long-term contexts to facilitate comprehensive understanding, and (2) improving data efficiency since constructing high-quality argument structures is time-consuming. In this work, we propose a novel context-aware Transformer-based argument structure prediction model which, on five different domains, significantly outperforms models that rely on features or only encode limited contexts. To tackle the difficulty of data annotation, we examine two complementary methods: (i) transfer learning to leverage existing annotated data to boost model performance in a new target domain, and (ii) active learning to strategically identify a small amount of samples for annotation. We further propose model-independent sample acquisition strategies, which can be generalized to diverse domains. With extensive experiments, we show that our simple-yet-effective acquisition strategies yield competitive results against three strong comparisons. Combined with transfer learning, substantial F1 score boost (5-25) can be further achieved during the early iterations of active learning across domains.
more »
« less
Boosting template-based SSVEP decoding by cross-domain transfer learning
Abstract Objective . This study aims to establish a generalized transfer-learning framework for boosting the performance of steady-state visual evoked potential (SSVEP)-based brain–computer interfaces (BCIs) by leveraging cross-domain data transferring. Approach . We enhanced the state-of-the-art template-based SSVEP decoding through incorporating a least-squares transformation (LST)-based transfer learning to leverage calibration data across multiple domains (sessions, subjects, and electroencephalogram montages). Main results . Study results verified the efficacy of LST in obviating the variability of SSVEPs when transferring existing data across domains. Furthermore, the LST-based method achieved significantly higher SSVEP-decoding accuracy than the standard task-related component analysis (TRCA)-based method and the non-LST naive transfer-learning method. Significance . This study demonstrated the capability of the LST-based transfer learning to leverage existing data across subjects and/or devices with an in-depth investigation of its rationale and behavior in various circumstances. The proposed framework significantly improved the SSVEP decoding accuracy over the standard TRCA approach when calibration data are limited. Its performance in calibration reduction could facilitate plug-and-play SSVEP-based BCIs and further practical applications.
more »
« less
- Award ID(s):
- 1935860
- PAR ID:
- 10341426
- Date Published:
- Journal Name:
- Journal of Neural Engineering
- Volume:
- 18
- Issue:
- 1
- ISSN:
- 1741-2560
- Page Range / eLocation ID:
- 016002
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract BackgroundThis research focused on the development of a motor imagery (MI) based brain–machine interface (BMI) using deep learning algorithms to control a lower-limb robotic exoskeleton. The study aimed to overcome the limitations of traditional BMI approaches by leveraging the advantages of deep learning, such as automated feature extraction and transfer learning. The experimental protocol to evaluate the BMI was designed as asynchronous, allowing subjects to perform mental tasks at their own will. MethodsA total of five healthy able-bodied subjects were enrolled in this study to participate in a series of experimental sessions. The brain signals from two of these sessions were used to develop a generic deep learning model through transfer learning. Subsequently, this model was fine-tuned during the remaining sessions and subjected to evaluation. Three distinct deep learning approaches were compared: one that did not undergo fine-tuning, another that fine-tuned all layers of the model, and a third one that fine-tuned only the last three layers. The evaluation phase involved the exclusive closed-loop control of the exoskeleton device by the participants’ neural activity using the second deep learning approach for the decoding. ResultsThe three deep learning approaches were assessed in comparison to an approach based on spatial features that was trained for each subject and experimental session, demonstrating their superior performance. Interestingly, the deep learning approach without fine-tuning achieved comparable performance to the features-based approach, indicating that a generic model trained on data from different individuals and previous sessions can yield similar efficacy. Among the three deep learning approaches compared, fine-tuning all layer weights demonstrated the highest performance. ConclusionThis research represents an initial stride toward future calibration-free methods. Despite the efforts to diminish calibration time by leveraging data from other subjects, complete elimination proved unattainable. The study’s discoveries hold notable significance for advancing calibration-free approaches, offering the promise of minimizing the need for training trials. Furthermore, the experimental evaluation protocol employed in this study aimed to replicate real-life scenarios, granting participants a higher degree of autonomy in decision-making regarding actions such as walking or stopping gait.more » « less
-
null (Ed.)Recent years have witnessed a growing body of research on autonomous activity recognition models for use in deployment of mobile systems in new settings such as when a wearable system is adopted by a new user. Current research, however, lacks comprehensive frameworks for transfer learning. Specifically, it lacks the ability to deal with partially available data in new settings. To address these limitations, we propose {\it OptiMapper}, a novel uninformed cross-subject transfer learning framework for activity recognition. OptiMapper is a combinatorial optimization framework that extracts abstract knowledge across subjects and utilizes this knowledge for developing a personalized and accurate activity recognition model in new subjects. To this end, a novel community-detection-based clustering of unlabeled data is proposed that uses the target user data to construct a network of unannotated sensor observations. The clusters of these target observations are then mapped onto the source clusters using a complete bipartite graph model. In the next step, the mapped labels are conditionally fused with the prediction of a base learner to create a personalized and labeled training dataset for the target user. We present two instantiations of OptiMapper. The first instantiation, which is applicable for transfer learning across domains with identical activity labels, performs a one-to-one bipartite mapping between clusters of the source and target users. The second instantiation performs optimal many-to-one mapping between the source clusters and those of the target. The many-to-one mapping allows us to find an optimal mapping even when the target dataset does not contain sufficient instances of all activity classes. We show that this type of cross-domain mapping can be formulated as a transportation problem and solved optimally. We evaluate our transfer learning techniques on several activity recognition datasets. Our results show that the proposed community detection approach can achieve, on average, 69%$ utilization of the datasets for clustering with an overall clustering accuracy of 87.5%. Our results also suggest that the proposed transfer learning algorithms can achieve up to 22.5% improvement in the activity recognition accuracy, compared to the state-of-the-art techniques. The experimental results also demonstrate high and sustained performance even in presence of partial data.more » « less
-
Abstract Objective.Non-invasive electroencephalograms (EEG)-based brain–computer interfaces (BCIs) play a crucial role in a diverse range of applications, including motor rehabilitation, assistive and communication technologies, holding potential promise to benefit users across various clinical spectrums. Effective integration of these applications into daily life requires systems that provide stable and reliable BCI control for extended periods. Our prior research introduced the AIRTrode, a self-adhesive (A), injectable (I), and room-temperature (RT) spontaneously-crosslinked hydrogel electrode (AIRTrode). The AIRTrode has shown lower skin-contact impedance and greater stability than dry electrodes and, unlike wet gel electrodes, does not dry out after just a few hours, enhancing its suitability for long-term application. This study aims to demonstrate the efficacy of AIRTrodes in facilitating reliable, stable and long-term online EEG-based BCI operations.Approach.In this study, four healthy participants utilized AIRTrodes in two BCI control tasks–continuous and discrete–across two sessions separated by six hours. Throughout this duration, the AIRTrodes remained attached to the participants’ heads. In the continuous task, participants controlled the BCI through decoding of upper-limb motor imagery (MI). In the discrete task, the control was based on decoding of error-related potentials (ErrPs).Main Results.Using AIRTrodes, participants demonstrated consistently reliable online BCI performance across both sessions and tasks. The physiological signals captured during MI and ErrPs tasks were valid and remained stable over sessions. Lastly, both the BCI performances and physiological signals captured were comparable with those from freshly applied, research-grade wet gel electrodes, the latter requiring inconvenient re-application at the start of the second session.Significance.AIRTrodes show great potential promise for integrating non-invasive BCIs into everyday settings due to their ability to support consistent BCI performances over extended periods. This technology could significantly enhance the usability of BCIs in real-world applications, facilitating continuous, all-day functionality that was previously challenging with existing electrode technologies.more » « less
-
null (Ed.)Large, comprehensive collections of single-cell RNA sequencing (scRNA-seq) datasets have been generated that allow for the full transcriptional characterization of cell types across a wide variety of biological and clinical conditions. As new methods arise to measure distinct cellular modalities, a key analytical challenge is to integrate these datasets or transfer knowledge from one to the other to better understand cellular identity and functions. Here, we present a simple yet surprisingly effective method named common factor integration and transfer learning (cFIT) for capturing various batch effects across experiments, technologies, subjects, and even species. The proposed method models the shared information between various datasets by a common factor space while allowing for unique distortions and shifts in genewise expression in each batch. The model parameters are learned under an iterative nonnegative matrix factorization (NMF) framework and then used for synchronized integration from across-domain assays. In addition, the model enables transferring via low-rank matrix from more informative data to allow for precise identification in data of lower quality. Compared with existing approaches, our method imposes weaker assumptions on the cell composition of each individual dataset; however, it is shown to be more reliable in preserving biological variations. We apply cFIT to multiple scRNA-seq datasets of developing brain from human and mouse, varying by technologies and developmental stages. The successful integration and transfer uncover the transcriptional resemblance across systems. The study helps establish a comprehensive landscape of brain cell-type diversity and provides insights into brain development.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1088/1741-2552/abcb6e
