Efficiently Train ASR Models that Memorize Less and Perform Better with Per-core Clipping
- URL: http://arxiv.org/abs/2406.02004v2
- Date: Wed, 5 Jun 2024 21:44:10 GMT
- Title: Efficiently Train ASR Models that Memorize Less and Perform Better with Per-core Clipping
- Authors: Lun Wang, Om Thakkar, Zhong Meng, Nicole Rafidi, Rohit Prabhavalkar, Arun Narayanan,
- Abstract summary: Per-core clip-ping (PCC) can effectively mitigate unintended memorization in ASR models.
PCC positively influences ASR performance metrics, leading to improved convergence rates and reduced word error rates.
- Score: 27.547461769425855
- License: http://creativecommons.org/licenses/by-sa/4.0/
- Abstract: Gradient clipping plays a vital role in training large-scale automatic speech recognition (ASR) models. It is typically applied to minibatch gradients to prevent gradient explosion, and to the individual sample gradients to mitigate unintended memorization. This work systematically investigates the impact of a specific granularity of gradient clipping, namely per-core clip-ping (PCC), across training a wide range of ASR models. We empirically demonstrate that PCC can effectively mitigate unintended memorization in ASR models. Surprisingly, we find that PCC positively influences ASR performance metrics, leading to improved convergence rates and reduced word error rates. To avoid tuning the additional hyperparameter introduced by PCC, we further propose a novel variant, adaptive per-core clipping (APCC), for streamlined optimization. Our findings highlight the multifaceted benefits of PCC as a strategy for robust, privacy-forward ASR model training.
Related papers
- Reactive Model Correction: Mitigating Harm to Task-Relevant Features via Conditional Bias Suppression [12.44857030152608]
Deep Neural Networks are prone to learning and relying on spurious correlations in the training data, which, for high-risk applications, can have fatal consequences.
Various approaches to suppress model reliance on harmful features have been proposed that can be applied post-hoc without additional training.
We propose a reactive approach conditioned on model-derived knowledge and eXplainable Artificial Intelligence (XAI) insights.
arXiv Detail & Related papers (2024-04-15T09:16:49Z) - Prior Constraints-based Reward Model Training for Aligning Large Language Models [58.33118716810208]
This paper proposes a Prior Constraints-based Reward Model (namely PCRM) training method to mitigate this problem.
PCRM incorporates prior constraints, specifically, length ratio and cosine similarity between outputs of each comparison pair, during reward model training to regulate optimization magnitude and control score margins.
Experimental results demonstrate that PCRM significantly improves alignment performance by effectively constraining reward score scaling.
arXiv Detail & Related papers (2024-04-01T07:49:11Z) - Model-Based Reparameterization Policy Gradient Methods: Theory and
Practical Algorithms [88.74308282658133]
Reization (RP) Policy Gradient Methods (PGMs) have been widely adopted for continuous control tasks in robotics and computer graphics.
Recent studies have revealed that, when applied to long-term reinforcement learning problems, model-based RP PGMs may experience chaotic and non-smooth optimization landscapes.
We propose a spectral normalization method to mitigate the exploding variance issue caused by long model unrolls.
arXiv Detail & Related papers (2023-10-30T18:43:21Z) - Unifying Synergies between Self-supervised Learning and Dynamic
Computation [53.66628188936682]
We present a novel perspective on the interplay between SSL and DC paradigms.
We show that it is feasible to simultaneously learn a dense and gated sub-network from scratch in a SSL setting.
The co-evolution during pre-training of both dense and gated encoder offers a good accuracy-efficiency trade-off.
arXiv Detail & Related papers (2023-01-22T17:12:58Z) - Online Convolutional Re-parameterization [51.97831675242173]
We present online convolutional re- parameterization (OREPA), a two-stage pipeline, aiming to reduce the huge training overhead by squeezing the complex training-time block into a single convolution.
Compared with the state-of-the-art re-param models, OREPA is able to save the training-time memory cost by about 70% and accelerate the training speed by around 2x.
We also conduct experiments on object detection and semantic segmentation and show consistent improvements on the downstream tasks.
arXiv Detail & Related papers (2022-04-02T09:50:19Z) - Learned Robust PCA: A Scalable Deep Unfolding Approach for
High-Dimensional Outlier Detection [23.687598836093333]
Robust principal component analysis is a critical tool in machine learning, which detects outliers in the task of low-rank reconstruction.
In this paper, we propose a scalable and learnable approach for high-dimensional RPCA problems which we call LRPCA.
We show that LRPCA outperforms the state-of-the-art RPCA algorithms, such as neurald AltProj, on both datasets real-world applications.
arXiv Detail & Related papers (2021-10-11T23:37:55Z) - Layer Pruning on Demand with Intermediate CTC [50.509073206630994]
We present a training and pruning method for ASR based on the connectionist temporal classification (CTC)
We show that a Transformer-CTC model can be pruned in various depth on demand, improving real-time factor from 0.005 to 0.002 on GPU.
arXiv Detail & Related papers (2021-06-17T02:40:18Z) - Extrapolation for Large-batch Training in Deep Learning [72.61259487233214]
We show that a host of variations can be covered in a unified framework that we propose.
We prove the convergence of this novel scheme and rigorously evaluate its empirical performance on ResNet, LSTM, and Transformer.
arXiv Detail & Related papers (2020-06-10T08:22:41Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.