skip to main content

Explore Research Products in the NSF-PAR

Title: An Investigation into the Contribution of Locally Aggregated Descriptors to Figurative Language Identification
In natural language understanding, topics that touch upon figurative language and pragmatics are notably difficult. We probe a novel use of locally aggregated descriptors -- specifically, an architecture called NeXtVLAD -- motivated by its accomplishments in computer vision, achieve tremendous success in the FigLang2020 sarcasm detection task. The reported F1 score of 93.1% is 14% higher than the next best result. We specifically investigate the extent to which the novel architecture is responsible for this boost, and find that it does not provide statistically significant benefits. Deep learning approaches are expensive, and we hope our insights highlighting the lack of benefits from introducing a resource-intensive component will aid future research to distill the effective elements from long and complex pipelines, thereby providing a boost to the wider research community.  more » « less
Award ID(s):
1834597
NSF-PAR ID:
10352069
Author(s) / Creator(s):
; ;
Date Published:
Journal Name:
Proceedings of the Second Workshop on Insights from Negative Results in NLP
Page Range / eLocation ID:
103 to 109
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ; ( , European Conference on Computer Vision)
    null (Ed.)
    Arguably one of the top success stories of deep learning is transfer learning. The finding that pre-training a network on a rich source set (e.g., ImageNet) can help boost performance once fine-tuned on a usually much smaller target set, has been instrumental to many applications in language and vision. Yet, very little is known about its usefulness in 3D point cloud understanding. We see this as an opportunity considering the effort required for annotating data in 3D. In this work, we aim at facilitating research on 3D representation learning. Different from previous works, we focus on high-level scene understanding tasks. To this end, we select a suit of diverse datasets and tasks to measure the effect of unsupervised pre-training on a large source set of 3D scenes. Our findings are extremely encouraging: using a unified triplet of architecture, source dataset, and contrastive loss for pre-training, we achieve improvement over recent best results in segmentation and detection across 6 different benchmarks for indoor and outdoor, real and synthetic datasets – demonstrating that the learned representation can generalize across domains. Furthermore, the improvement was similar to supervised pre-training, suggesting that future efforts should favor scaling data collection over more detailed annotation. We hope these findings will encourage more research on unsupervised pretext task design for 3D deep learning. 
    more » « less
  2. ; ; ; ( , IEEE Global Conference on Signal and Information Processing (GlobalSIP))
    With benefits of fast query speed and low storage cost, hashing-based image retrieval approaches have garnered considerable attention from the research community. In this paper, we propose a novel Error-Corrected Deep Cross Modal Hashing (CMH-ECC) method which uses a bitmap specifying the presence of certain facial attributes as an input query to retrieve relevant face images from the database. In this architecture, we generate compact hash codes using an end-to-end deep learning module, which effectively captures the inherent relationships between the face and attribute modality. We also integrate our deep learning module with forward error correction codes to further reduce the distance between different modalities of the same subject. Specifically, the properties of deep hashing and forward error correction codes are exploited to design a cross modal hashing framework with high retrieval performance. Experimental results using two standard datasets with facial attributes-image modalities indicate that our CMH-ECC face image retrieval model outperforms most of the current attribute-based face image retrieval approaches. 
    more » « less
  3. ; ; ; ; ; ; ; ; ( , The International Network for the Science of Team Science (INSciTS))
    Broadband infrastructure in urban parks may serve crucial functions including an amenity to boost overall park use and a bridge to propagate WiFi access into contiguous neighborhoods. This project: SCC:PG Park WiFi as a BRIDGE to Community Resilience has developed a new model —Build Resilience through the Internet and Digital Greenspace Exposure, leveraging off-the-shelf WiFi technology, novel algorithms, community assets, and local partnerships to lower greenspace WiFi costs. This interdisciplinary work leverages: computer science, information studies, landscape architecture, and public health. Collaboration methodologies and relational definitions across disciplines are still nascent —especially when paired with civic-engaged, applied research. Student researchers (UG/Grad) are excellent partners in bridging disciplinary barriers and constraints. Their capacity to assimilate multiple frameworks has produced refinements to the project’s theoretical lenses and suggested novel socio-technical methodology improvements. Further, they are excellent ambassadors to community partners and stakeholders. In BRIDGE, we tested two mechanisms to augment student research participation. In both, we leveraged a classic, curriculum-based model named the Partnership for Action Learning in Sustainability program (PALS). This campus-wide, community-engaged initiative pairs faculty and students with community partners. PALS curates economic, environmental, and social sustainability challenges and scopes projects to customize appropriate coursework that addresses identified challenges. Outcomes include: literature searches, wireframes, and design plans that target solutions to civic problems. Constraints include the short semester timeframe and curriculum-learning-outcome constraints. (1) On BRIDGE, Dr. Kweon executed a semester-based Landscape Architecture PALS 400-level-studio. 18 undergraduates conducted in-class and in-field work to assess community needs and proposed design solutions for future park-wide WiFi. Research topics included: community-park history, neighborhood demographics, case-study analysis, and land-cover characteristics. The students conducted an in-Park, community engagement session —via interactive posterboard surveys, to gain input on what park amenities might be redesigned or added to promote WiFi use. The students then produced seven re-design plans; one included a café/garden, with an eco-corridor that integrated technology with nature. (2) From the classic, curriculum-based PALS model we created a summer-intensive for our five research assistants, to stimulate interdisciplinary collaboration in their research tasks and co-analysis of project data products: experimental technical WiFi-setup, community survey results, and stakeholder needs-assessments. Students met weekly with each other and team leadership, exchanged journal articles, and attended joint research events. This model shows promise for integrating students more formally into an interdisciplinary research project. An end-of-intensive focus group highlighted, from the students’ perspective, the pro/cons of this model. Results: In contrasting the two mechanisms, our results include: Model 1 is tried-and-trued and produces standardized, reliable products. However, as work is group based, student independence is limited —to explore topics/themes of interest. Civic groups are typically thrilled with the diversity of action plans produced. Model 2 provides greater independence in student-learning outcomes, fosters interdisciplinary, “dictionary-building” that can be used by the full team, deepens methodological approaches, and allows for student stipend payments. Lessons learned: intensive time frame needed more research team support and ideally should be extended, when possible, over the full project-span. UMD-IRB#1785365-4; NSF-award: 2125526. 
    more » « less
  4. ; ; ; ; ; ( , Advances in neural information processing systems)
    Transfer learning from the model trained on large datasets to customized downstream tasks has been widely used as the pre-trained model can greatly boost the generalizability. However, the increasing sizes of pre-trained models also lead to a prohibitively large memory footprints for downstream transferring, making them unaffordable for personal devices. Previous work recognizes the bottleneck of the footprint to be the activation, and hence proposes various solutions such as injecting specific lite modules. In this work, we present a novel memory-efficient transfer framework called Back Razor, that can be plug-and-play applied to any pre-trained network without changing its architecture. The key idea of Back Razor is asymmetric sparsifying: pruning the activation stored for back-propagation, while keeping the forward activation dense. It is based on the observation that the stored activation, that dominates the memory footprint, is only needed for backpropagation. Such asymmetric pruning avoids affecting the precision of forward computation, thus making more aggressive pruning possible. Furthermore, we conduct the theoretical analysis for the convergence rate of Back Razor, showing that under mild conditions, our method retains the similar convergence rate as vanilla SGD. Extensive transfer learning experiments on both Convolutional Neural Networks and Vision Transformers with classification, dense prediction, and language modeling tasks show that Back Razor could yield up to 97% sparsity, saving 9.2x memory usage, without losing accuracy. The code is available at: https://github.com/VITA-Group/BackRazor_Neurips22. 
    more » « less
  5. ; ; ; ( , 1st Workshop on Natural Language Processing for Programming (NLP4Prog))
    Decompiling binary executables to high-level code is an important step in reverse engineering scenarios, such as malware analysis and legacy code maintenance. However, the generated high-level code is difficult to understand since the original variable names are lost. In this paper, we leverage transformer models to reconstruct the original variable names from decompiled code. Inherent differences between code and natural language present certain challenges in applying conventional transformer-based architectures to variable name recovery. We propose DIRECT, a novel transformer-based architecture customized specifically for the task at hand. We evaluate our model on a dataset of decompiled functions and find that DIRECT outperforms the previous state-of-the-art model by up to 20%. We also present ablation studies evaluating the impact of each of our modifications. We make the source code of DIRECT available to encourage reproducible research. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email