Enhancing Speaker Diarization with Large Language Models: A Contextual Beam Search Approach [PDF]
IEEE International Conference on Acoustics, Speech, and Signal Processing, 2023 Large language models (LLMs) have shown great promise for capturing contextual information in natural language processing tasks. We propose a novel approach to speaker diarization that incorporates the prowess of LLMs to exploit contextual cues in human ...
T. Park +3 more
semanticscholar +1 more source
Conformal Autoregressive Generation: Beam Search with Coverage Guarantees [PDF]
AAAI Conference on Artificial Intelligence, 2023 We introduce two new extensions to the beam search algorithm based on conformal predictions (CP) to produce sets of sequences with theoretical coverage guarantees.
Nicolas Deutschmann +2 more
semanticscholar +1 more source
On the Feasibility of Out-of-Band Spatial Channel Information for Millimeter-Wave Beam Search [PDF]
IEEE Transactions on Antennas and Propagation, 2022 Prolonged beam alignment is the main source of overhead in mobile wireless communications at millimeter-wave (mm-wave) frequencies due to narrow beams following the requirement of high antenna gains. Out-of-band spatial information may be used in initial
Peize Zhang +4 more
semanticscholar +1 more source
Streaming parallel transducer beam search with fast-slow cascaded encoders [PDF]
Interspeech, 2022 Streaming ASR with strict latency constraints is required in many speech recognition applications. In order to achieve the required latency, streaming ASR models sacrifice accuracy compared to non-streaming ASR models due to lack of future input context.
Jay Mahadeokar +7 more
semanticscholar +1 more source
Improved Beam Search for Hallucination Mitigation in Abstractive Summarization [PDF]
arXiv.org, 2022 Advancement in large pretrained language models has significantly improved their performance for conditional language generation tasks including summarization albeit with hallucinations.
A. Sridhar, Erik M. Visser
semanticscholar +1 more source
SentBS: Sentence-level Beam Search for Controllable Summarization [PDF]
Conference on Empirical Methods in Natural Language Processing, 2022 A wide range of control perspectives have been explored in controllable text generation. Structure-controlled summarization is recently proposed as a useful and interesting research direction.
Chenhui Shen +4 more
semanticscholar +1 more source
BREAK: Breaking the Dialogue State Tracking Barrier with Beam Search and Re-ranking
Annual Meeting of the Association for Computational Linguistics, 2023 Despite the recent advances in dialogue state tracking (DST), the joint goal accuracy (JGA) of the existing methods on MultiWOZ 2.1 still remains merely 60%.
Seungpil Won +4 more
semanticscholar +1 more source
First the Worst: Finding Better Gender Translations During Beam Search [PDF]
Findings, 2021 Generating machine translations via beam search seeks the most likely output under a model. However, beam search has been shown to amplify demographic biases exhibited by a model.
D. Saunders, Rosie Sallis, B. Byrne
semanticscholar +1 more source
Neural Network Approach to Program Synthesis for Tabular Transformation by Example
IEEE Access, 2022 Data transformation is a laborious and time-consuming task for analysts. Programming by example (PBE) is a technique that can simplify this difficult task for data analysts by automatically generating programs for data transformation.
Yoshifumi Ujibashi, Atsuhiro Takasu
doaj +1 more source
LinearFold: linear-time approximate RNA folding by 5'-to-3' dynamic programming and beam search. [PDF]
Bioinformatics, 2019 Motivation Predicting the secondary structure of an ribonucleic acid (RNA) sequence is useful in many applications. Existing algorithms [based on dynamic programming] suffer from a major limitation: their runtimes scale cubically with the RNA length, and
Huang L +6 more
europepmc +3 more sources