Publications (21)
A dataset for resolving referring expressions in spoken dialogue via contextual query rewrites (CQR)
Michael Regan, Pushpendre Rastogi, Arpit Gupta +1
We present Contextual Query Rewrite (CQR) a dataset for multi-domain task-oriented spoken dialogue systems that is an extension of the Stanford dialog corpus (Eric et al., 2017a).…
Meta-training with Demonstration Retrieval for Efficient Few-shot Learning
Aaron Mueller, Kanika Narang, Lambert Mathias +2
Large language models show impressive results on few-shot NLP tasks. However, these models are memory and computation-intensive. Meta-training allows one to leverage smaller models…
Time Masking: Leveraging Temporal Information in Spoken Dialogue Systems
Rylan Conway, Lambert Mathias
In a spoken dialogue system, dialogue state tracker (DST) components track the state of the conversation by updating a distribution of values associated with each of the slots bein…
Pixel-Grounded Retrieval for Knowledgeable Large Multimodal Models
Jeonghwan Kim, Renjie Tao, Sanat Sharma +8
Visual Question Answering (VQA) often requires coupling fine-grained perception with factual knowledge beyond the input image. Prior multimodal Retrieval-Augmented Generation (MM-R…
Logical Satisfiability of Counterfactuals for Faithful Explanations in NLI
Suzanna Sia, Anton Belyy, Amjad Almahairi +3
Evaluating an explanation's faithfulness is desired for many reasons such as trust, interpretability and diagnosing the sources of model's errors. In this work, which focuses on th…
Policy Compliance Detection via Expression Tree Inference
Neema Kotonya, Andreas Vlachos, Majid Yazdani +2
Policy Compliance Detection (PCD) is a task we encounter when reasoning over texts, e.g. legal frameworks. Previous work to address PCD relies heavily on modeling the task as a spe…
Pre-Training for Query Rewriting in A Spoken Language Understanding System
Zheng Chen, Xing Fan, Yuan Ling +2
Query rewriting (QR) is an increasingly important technique to reduce customer friction caused by errors in a spoken language understanding pipeline, where the errors originate fro…
Scaling Multi-Domain Dialogue State Tracking via Query Reformulation
Pushpendre Rastogi, Arpit Gupta, Tongfei Chen +1
We present a novel approach to dialogue state tracking and referring expression resolution tasks. Successful contextual understanding of multi-turn spoken dialogues requires resolv…
UniPELT: A Unified Framework for Parameter-Efficient Language Model Tuning
Yuning Mao, Lambert Mathias, Rui Hou +5
Recent parameter-efficient language model tuning (PELT) methods manage to match the performance of fine-tuning with much fewer trainable parameters and perform especially well when…
ToKen: Task Decomposition and Knowledge Infusion for Few-Shot Hate Speech Detection
Badr AlKhamissi, Faisal Ladhak, Srini Iyer +7
Hate speech detection is complex; it relies on commonsense reasoning, knowledge of stereotypes, and an understanding of social nuance that differs from one culture to the next. It…
Improving Long Distance Slot Carryover in Spoken Dialogue Systems
Tongfei Chen, Chetan Naik, Hua He +2
Tracking the state of the conversation is a central component in task-oriented spoken dialogue systems. One such approach for tracking the dialogue state is slot carryover, where a…
Transfer Learning for Neural Semantic Parsing
Xing Fan, Emilio Monti, Lambert Mathias +1
The goal of semantic parsing is to map natural language to a machine interpretable meaning representation language (MRL). One of the constraints that limits full exploration of dee…
Leveraging External Knowledge for Out-Of-Vocabulary Entity Labeling
Adrian de Wynter, Lambert Mathias
Dealing with previously unseen slots is a challenging problem in a real-world multi-domain dialogue state tracking task. Other approaches rely on predefined mappings to generate ca…
Contextual Slot Carryover for Disparate Schemas
Chetan Naik, Arpit Gupta, Hancheng Ge +2
In the slot-filling paradigm, where a user can refer back to slots in the context during a conversation, the goal of the contextual understanding system is to resolve the referring…
PERFECT: Prompt-free and Efficient Few-shot Learning with Language Models
Rabeeh Karimi Mahabadi, Luke Zettlemoyer, James Henderson +4
Current methods for few-shot fine-tuning of pretrained masked language models (PLMs) require carefully engineered prompts and verbalizers for each new task to convert examples into…
Personalized Query Rewriting in Conversational AI Agents
Alireza Roshan-Ghias, Clint Solomon Mathialagan, Pragaash Ponnusamy +2
Spoken language understanding (SLU) systems in conversational AI agents often experience errors in the form of misrecognitions by automatic speech recognition (ASR) or semantic gap…
TRACE: A Framework for Analyzing and Enhancing Stepwise Reasoning in Vision-Language Models
Shima Imani, Seungwhan Moon, Lambert Mathias +2
Reliable mathematical and scientific reasoning remains an open challenge for large vision-language models. Standard final-answer evaluation often masks reasoning errors, allowing s…
Reading Recognition in the Wild
Charig Yang, Samiul Alam, Shakhrul Iman Siam +12
To enable egocentric contextual AI in always-on smart glasses, it is crucial to be able to keep a record of the user's interactions with the world, including during reading. In thi…
Cross-Lingual Approaches to Reference Resolution in Dialogue Systems
Amr Sharaf, Arpit Gupta, Hancheng Ge +2
In the slot-filling paradigm, where a user can refer back to slots in the context during the conversation, the goal of the contextual understanding system is to resolve the referri…
TimelineQA: A Benchmark for Question Answering over Timelines
Wang-Chiew Tan, Jane Dwivedi-Yu, Yuliang Li +4
Lifelogs are descriptions of experiences that a person had during their life. Lifelogs are created by fusing data from the multitude of digital services, such as online photos, map…
UNIREX: A Unified Learning Framework for Language Model Rationale Extraction
Aaron Chan, Maziar Sanjabi, Lambert Mathias +5
An extractive rationale explains a language model's (LM's) prediction on a given task instance by highlighting the text inputs that most influenced the prediction. Ideally, rationa…