On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

169,181 papers · 148 categories

Trend · papers per month

count % of corpus

11 results for “TDNNs”

Improved TDNNs using Deep Kernels and Frequency Dependent Grid-RNNscs.CL

Improved TDNNs enhance speech recognition with deep kernels and frequency-dependent processing.

problem Shallow TDNN models limit long-term context modeling.

method Deepened kernels with residual connections and spectro-temporal processing.

result Deep kernel TDNNs reduce WER by 6% and further by 9% with frequency-dependent Grid-RNN.

TDNNs speech recognition deep learning residual connections spectro-temporal

2018-02-18abs ↗pdf ↗

Semi-supervised acoustic and language model training for English-isiZulu code-switched speech recognitioneess.AS

Improved ASR for English-isiZulu code-switched speech with semi-supervised training.

problem Improving ASR for code-switched speech between English and isiZulu.

method Semi-supervised training using automatic transcription of multilingual speech data.

result Semi-supervised training achieved significant WER reduction in ASR performance.

code-switched speech semi-supervised training ASR TDNN-F CNN

2020-04-05abs ↗pdf ↗

MLP, XGBoost, KAN, TDNN, and LSTM-GRU Hybrid RNN with Attention for SPX and NDX European Call Option Pricingq-fin.CP

Study compares various ANN models for SPX and NDX options pricing.

problem Approximating complex multivariate functions for accurate option pricing.

method Hybrid RNN models (LSTM-GRU, TDNN, MLP, KAN) with attention mechanisms.

result LSTM-GRU hybrid RNN with attention outperforms other models.

ANN RNN Option Pricing Black-Scholes Supervised Learning

2024-08-26abs ↗pdf ↗

Bayesian Learning for Deep Neural Network Adaptationcs.SD

Bayesian approach improves speech recognition with limited speaker data.

problem Reduces mismatch between training and evaluation data due to speaker differences.

method Bayesian learning for DNN adaptation models with limited speaker data.

result Bayesian adaptation consistently outperforms deterministic methods, reducing word error rates up to 1.4%.

Bayesian learning Deep neural networks Speaker adaptation Limited data Speech recognition

2020-12-14abs ↗pdf ↗

Finite element form-valued forms: Constructionmath.NA

New finite element method for complex forms in any dimension.

problem Discretization of complex forms in arbitrary dimensions.

method Finite element discretization of

\ell

-form-valued

k

-forms on triangulations.

result Generalizes existing finite element methods for various tensor fields.

finite elements form-valued forms discretization continuum mechanics differential geometry

2025-03-05abs ↗pdf ↗

A Comparison of Lattice-free Discriminative Training Criteria for Purely Sequence-Trained Neural Network Acoustic Modelscs.LG

Three LF training criteria improve neural network acoustic models without cross-entropy pre-training.

problem Improving purely sequence-trained neural network acoustic models.

method Comparison of three lattice-free discriminative training criteria (MMI, bMMI, sMBR) on LVCSR tasks.

result LF-bMMI models outperform plain LF-MMI models by 5% WER on Switchboard datasets.

neural network acoustic models lattice-free training sequence training maximum mutual information boosted maximum mutual information

2018-11-08abs ↗pdf ↗

Automatic Speech Recognition for Humanitarian Applications in Somalics.CL

System builds Somali ASR for UN humanitarian efforts.

problem Developing ASR for under-resourced Somali language.

method Acoustic model training with annotated speech, neural architectures, language model data augmentation, acoustic data perturbation.

result Best system achieved 53.75% word error rate.

Somali ASR Humanitarian Acoustic Model Language Model Data Augmentation

2018-07-23abs ↗pdf ↗

A Comparison of Hybrid and End-to-End Models for Syllable Recognitioncs.CL

Hybrid and end-to-end models compare in syllable recognition.

problem Comparing hybrid and end-to-end models for syllable recognition.

method Traditional hybrid system (kaldi) vs. end-to-end (TensorFlow) models.

result Hybrid models with explicit syllable knowledge outperform end-to-end models.

syllable recognition end-to-end models hybrid models German Verbmobil corpus

2019-09-19abs ↗pdf ↗

Deep Speaker Embeddings for Far-Field Speaker Recognition on Short Utterancescs.SD

Improved far-field speaker verification for short utterances in noisy conditions.

problem Challenges in speaker verification on short utterances in uncontrolled noisy environments.

method Used deep neural network architectures (TDNN and ResNet) and experimented with various embedding extractors and training procedures.

result ResNet architectures outperform x-vector approach in speaker verification quality for both long and short utterances.

speaker recognition far-field short utterances residual networks embedding extractors

2020-02-14abs ↗pdf ↗

Joint Binary Neural Network for Multi-label Learning with Applications to Emotion Classificationcs.LG

Proposes JBNN for multi-label classification with improved efficiency and performance.

problem Multi-label classification with dependencies and heavy computational load.

method Joint Binary Neural Network (JBNN) that synchronously performs multiple binary classifications and captures label relations via joint binary cross entropy (JBCE) loss.

result Significantly better performance and computational efficiency compared to state-of-the-art methods.

multi-label classification binary neural network joint binary cross entropy emotion classification prior label relations

2018-02-03abs ↗pdf ↗

Optimal Nonparametric Inference with Two-Scale Distributional Nearest Neighborsstat.ML

A new method improves the DNN estimator for mean regression, achieving optimal convergence rates.

problem Lack of distributional results and suboptimal convergence rates for DNN estimator.

method Two-scale distributional nearest neighbors (TDNN) estimator, combining two DNN estimators with different subsampling scales.

result The two-scale DNN estimator achieves the optimal nonparametric convergence rate under fourth-order smoothness.

nonparametric inference nearest neighbors bias reduction distributional results optimal convergence rate

2018-08-25abs ↗pdf ↗