A5084241804- Topic Modeling
- Natural Language Processing Techniques
- Speech and dialogue systems
- Multimodal Machine Learning Applications
- Innovative Teaching and Learning Methods
- EEG and Brain-Computer Interfaces
- Explainable Artificial Intelligence (XAI)
- Neuroscience and Neural Engineering
- Intelligent Tutoring Systems and Adaptive Learning
- Text Readability and Simplification
- Algorithms and Data Compression
- Adversarial Robustness in Machine Learning
- Machine Learning in Healthcare
- Knowledge Management and Sharing
- Advanced Memory and Neural Computing
- Human Pose and Action Recognition
- Advanced Text Analysis Techniques
- Sepsis Diagnosis and Treatment
- Muscle activation and electromyography studies
- Video Analysis and Summarization
- Online Learning and Analytics
- Sex work and related issues
- Software Engineering Techniques and Practices
- Music and Audio Processing
- Emergency and Acute Care Studies
Palo Alto University
2024
Stanford University
2024
University of Colorado System
2021-2023
University of Colorado Boulder
2021-2023
University of Colorado Denver
2022
Neural Signals (United States)
2022
University of Massachusetts Amherst
2018-2021
Recent progress in hardware and methodology for training neural networks has ushered a new generation of large trained on abundant data. These models have obtained notable gains accuracy across many NLP tasks. However, these improvements depend the availability exceptionally computational resources that necessitate similarly substantial energy consumption. As result are costly to train develop, both financially, due cost electricity or cloud compute time, environmentally, carbon footprint...
The field of artificial intelligence has experienced a dramatic methodological shift towards large neural networks trained on plentiful data. This been fueled by recent advances in hardware and techniques enabling remarkable levels computation, resulting impressive AI across many applications. However, the massive computation required to obtain these exciting results is costly both financially, due price specialized electricity or cloud compute time, environment, as result non-renewable...
Recent progress in hardware and methodology for training neural networks has ushered a new generation of large trained on abundant data. These models have obtained notable gains accuracy across many NLP tasks. However, these improvements depend the availability exceptionally computational resources that necessitate similarly substantial energy consumption. As result are costly to train develop, both financially, due cost electricity or cloud compute time, environmentally, carbon footprint...
The use of Artificial Intelligence (AI) in K-12 education is showing considerable promise to enhance student learning, yet existing tools continue situate AI tutoring firmly within the context one-on-one instruction and personalized learning. As HCI, learning science, team science researchers we envision help students become better collaborators—a highly valued skill for their lives after school. In this demonstration present "CoBi"—a multi-party partner that focuses on relationship...
In collaborative learning environments, effective intelligent systems need to accurately analyze and understand the discourse between learners (i.e., group modeling) provide adaptive support. We investigate how automatic speech recognition (ASR) errors influence models of small collaboration in noisy real-world classrooms. Our dataset consisted 30 students recorded by consumer off-the-shelf microphones (Yeti Blue) while engaging dyadic- triadic- a multi-day STEM curriculum unit. found that...
Legal documents such as contracts contain complex and domain-specific jargons, long nested sentences, often present with several details that may be difficult to understand for laypeople without domain expertise.In this paper, we explore the problem of text simplification (TS) in legal domain.The main challenge is lack availability complex-simple parallel datasets domain.We investigate some existing datasets, methods, metrics TS literature simplifying texts, perform human evaluation analyze...
Haw-Shiuan Chang, Amol Agrawal, Ananya Ganesh, Anirudha Desai, Vinayak Mathur, Alfred Hough, Andrew McCallum. Proceedings of the Twelfth Workshop on Graph-Based Methods for Natural Language Processing (TextGraphs-12). 2018.
Advances in conversational AI systems, powered particular by large language models, have facilitated rapid progress understanding and generating dialog. Typically, task-oriented or open-domain dialog systems been designed to work with two-party dialog, i.e., the exchange of utterances between a single user system. However, modern may be deployed scenarios such as classrooms meetings where analysis multiple speakers is required. This survey will present research around computational modeling...
Ananya Ganesh, Hugh Scribner, Jasdeep Singh, Katherine Goodman, Jean Hertzberg, Katharina Kann. Proceedings of the 17th Workshop on Innovative Use NLP for Building Educational Applications (BEA 2022). 2022.
The motivation of someone who is locked-in, that is, paralyzed and mute, to find relief for their loss function. data presented in this report part an attempt restore one those lost functions, namely, speech. An essential feature the development a speech prosthesis optimal decoding patterns recorded neural signals during silent or covert speech, speaking "inside head" with output inaudible due paralysis articulators. aim paper illustrate importance both fast slow single unit firings from...
Recent advances in natural language processing (NLP) have the ability to transform how classroom learning takes place.Combined with increasing integration of technology today's classrooms, NLP systems leveraging question answering and dialog techniques can serve as private tutors or participants discussions increase student engagement learning.To progress towards this goal, we use discourse framework academically productive talk (APT) learn strategies that make for best experience.In paper,...
In this paper, we introduce Dependency Dialogue Acts (DDA), a novel framework for capturing the structure of speaker-intentions in multi-party dialogues. DDA combines and adapts features from existing dialogue annotation frameworks, emphasizes multi-relational response dialogues addition to acts rhetorical relations. It represents functional, discourse, multi-threaded conversations. A few key distinguish frameworks such as SWBD-DAMSL ISO 24617-2 standard. First, prioritizes relational units...
Ananya Ganesh, Jie Cao, E. Margaret Perkoff, Rosy Southwell, Martha Palmer, Katharina Kann. Proceedings of the 61st Annual Meeting Association for Computational Linguistics (Volume 2: Short Papers). 2023.
In this study, we investigate the performance of computer vision AI algorithms in predicting patient disposition from emergency department (ED) using short video clips. Clinicians often use "eye-balling" or clinical gestalt to aid triage, based on brief observations. We hypothesize that can similarly appearance for prediction. Data were collected adult patients at an academic ED, with mobile phone videos capturing performing simple tasks. Our algorithm, alone, showed better hospital...
Word sense induction (WSI), which addresses polysemy by unsupervised discovery of multiple word senses, resolves ambiguities for downstream NLP tasks and also makes representations more interpretable. This paper proposes an accurate efficient graph-based method WSI that builds a global non-negative vector embedding basis (which are interpretable like topics) clusters the indexes in ego network each polysemous word. By adopting distributional inclusion embeddings as our formation model, we...
Abstract Summary The motivation of someone who is locked-in, that is, paralyzed and mute, to find relief for their loss function. data presented in this report part an attempt restore one those lost functions, namely, speech. An essential feature the development a speech prosthetic optimal decoding patterns recorded neural signals during silent or covert speech, speaking ‘inside head’ with no audible output due paralysis articulators. aim paper illustrate importance both fast slow single...
Neural machine translation (MT) systems have been shown to perform poorly on low-resource language pairs, for which large-scale parallel data is unavailable. Making the annotation process faster and cheaper therefore important ensure equitable access MT systems. To make optimal use of a limited budget, we present CHIA (choosing instances annotate), method selecting annotate translation. Using an existing multi-way dataset high-resource languages, first identify instances, based model...
Rajat Bhatnagar, Ananya Ganesh, Katharina Kann. Proceedings of the 59th Annual Meeting Association for Computational Linguistics and 11th International Joint Conference on Natural Language Processing (Volume 2: Short Papers). 2021.