Search Results for author:
Found 15 papers, 0 papers with code
Weight subcloning: direct initialization of transformers using larger pretrained ones
The usual practice of transfer learning overcomes this challenge by initializing the model with weights of a pretrained model of the same size and specification to increase the convergence and training speed.
eDKM: An Efficient and Accurate Train-time Weight Clustering for Large Language Models
Since Large Language Models or LLMs have demonstrated high-quality performance on many complex language tasks, there is a great interest in bringing these LLMs to mobile devices for faster responses and better privacy protection.
Improving vision-inspired keyword spotting using dynamic module skipping in streaming conformer encoder
Using a vision-inspired keyword spotting framework, we propose an architecture with input-dependent dynamic depth capable of processing streaming audio.
Flexible Keyword Spotting based on Homogeneous Audio-Text Embedding
Spotting user-defined/flexible keywords represented in text frequently uses an expensive text encoder for joint analysis with an audio encoder in an embedding space, which can suffer from heterogeneous modality representation (i. e., large mismatch) and increased complexity.
Matching Latent Encoding for Audio-Text based Keyword Spotting
Using audio and text embeddings jointly for Keyword Spotting (KWS) has shown high-quality results, but the key challenge of how to semantically align two embeddings for multi-word keywords of different sequence lengths remains largely unsolved.
HEiMDaL: Highly Efficient Method for Detection and Localization of wake-words
Streaming keyword spotting is a widely used solution for activating voice assistants.
Ranked #1 on
Keyword Spotting
on hey Siri
I see what you hear: a vision-inspired method to localize words
This paper explores the possibility of using visual object detection techniques for word localization in speech data.
Optimize what matters: Training DNN-HMM Keyword Spotting Model Using End Metric
The DNN, in prior methods, is trained independent of the HMM parameters to minimize the cross-entropy loss between the predicted and the ground-truth state probabilities.
Ranked #2 on
Keyword Spotting
on hey Siri
Knowledge Transfer for Efficient On-device False Trigger Mitigation
But in case of a false trigger, transcribing the audio using ASR itself is strongly undesirable.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+2
Complementary Language Model and Parallel Bi-LRNN for False Trigger Mitigation
In this paper, we propose a novel solution to the FTM problem by introducing a parallel ASR decoding process with a special language model trained from "out-of-domain" data sources.
On the Role of Visual Cues in Audiovisual Speech Enhancement
One byproduct of this finding is that the learned visual embeddings can be used as features for other visual speech applications.
Detecting Emotion Primitives from Speech and their use in discerning Categorical Emotions
Emotion plays an essential role in human-to-human communication, enabling us to convey feelings such as happiness, frustration, and sincerity.
Multi-task Learning for Speaker Verification and Voice Trigger Detection
We train the network in a supervised multi-task learning setup, where the speech transcription branch of the network is trained to minimise a phonetic connectionist temporal classification (CTC) loss while the speaker recognition branch of the network is trained to label the input sequence with the correct label for the speaker.
Lattice-based Improvements for Voice Triggering Using Graph Neural Networks
A pure trigger-phrase detector model doesn't fully utilize the intent of the user speech whereas by using the complete decoding lattice of user audio, we can effectively mitigate speech not intended for the smart assistant.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+1
Leveraging Acoustic Cues and Paralinguistic Embeddings to Detect Expression from Voice
The expectation is that such assistants should understand the intent of the users query.
