This paper presents an extension to train end-to-end Context-Aware Transformer Transducer ( CATT ) models by using a simple, yet efficient method of mining hard negative phrases from the latent space of the context encoder. During training, given a reference query, we mine a number of similar phrases using approximate nearest neighbour search. These sampled phrases are then used as negative examples in the context list alongside random and ground truth contextual information. By including approximate nearest neighbour phrases (ANN-P) in the context list, we encourage the learned representation to disambiguate between similar, but not identical, biasing phrases. This improves biasing accuracy when there are several similar phrases in the biasing inventory. We carry out experiments in a large-scale data regime obtaining up to 7% relative word error rate reductions for the contextual portion of test data. We also extend and evaluate CATT approach in streaming applications.
相關內容
Large multimodal models (LMM) have recently shown encouraging progress with visual instruction tuning. In this note, we show that the fully-connected vision-language cross-modal connector in LLaVA is surprisingly powerful and data-efficient. With simple modifications to LLaVA, namely, using CLIP-ViT-L-336px with an MLP projection and adding academic-task-oriented VQA data with simple response formatting prompts, we establish stronger baselines that achieve state-of-the-art across 11 benchmarks. Our final 13B checkpoint uses merely 1.2M publicly available data, and finishes full training in ~1 day on a single 8-A100 node. We hope this can make state-of-the-art LMM research more accessible. Code and model will be publicly available.
Large-scale pre-trained Vision Language Models (VLMs) have proven effective for zero-shot classification. Despite the success, most traditional VLMs-based methods are restricted by the assumption of partial source supervision or ideal vocabularies, which rarely satisfy the open-world scenario. In this paper, we aim at a more challenging setting, Realistic Zero-Shot Classification, which assumes no annotation but instead a broad vocabulary. To address this challenge, we propose the Self Structural Semantic Alignment (S^3A) framework, which extracts the structural semantic information from unlabeled data while simultaneously self-learning. Our S^3A framework adopts a unique Cluster-Vote-Prompt-Realign (CVPR) algorithm, which iteratively groups unlabeled data to derive structural semantics for pseudo-supervision. Our CVPR process includes iterative clustering on images, voting within each cluster to identify initial class candidates from the vocabulary, generating discriminative prompts with large language models to discern confusing candidates, and realigning images and the vocabulary as structural semantic alignment. Finally, we propose to self-learn the CLIP image encoder with both individual and structural semantic alignment through a teacher-student learning strategy. Our comprehensive experiments across various generic and fine-grained benchmarks demonstrate that the S^3A method offers substantial improvements over existing VLMs-based approaches, achieving a more than 15% accuracy improvement over CLIP on average. Our codes, models, and prompts are publicly released at //github.com/sheng-eatamath/S3A.
Artificial Neural Networks (ANNs) trained with Backpropagation (BP) show astounding performance and are increasingly often used in performing our daily life tasks. However, ANNs are highly vulnerable to adversarial attacks, which alter inputs with small targeted perturbations that drastically disrupt the models' performance. The most effective method to make ANNs robust against these attacks is adversarial training, in which the training dataset is augmented with exemplary adversarial samples. Unfortunately, this approach has the drawback of increased training complexity since generating adversarial samples is very computationally demanding. In contrast to ANNs, humans are not susceptible to adversarial attacks. Therefore, in this work, we investigate whether biologically-plausible learning algorithms are more robust against adversarial attacks than BP. In particular, we present an extensive comparative analysis of the adversarial robustness of BP and Present the Error to Perturb the Input To modulate Activity (PEPITA), a recently proposed biologically-plausible learning algorithm, on various computer vision tasks. We observe that PEPITA has higher intrinsic adversarial robustness and, with adversarial training, has a more favourable natural-vs-adversarial performance trade-off as, for the same natural accuracies, PEPITA's adversarial accuracies decrease in average by 0.26% and BP's by 8.05%.
State-of-the-art sequential recommendation relies heavily on self-attention-based recommender models. Yet such models are computationally expensive and often too slow for real-time recommendation. Furthermore, the self-attention operation is performed at a sequence-level, thereby making low-cost incremental inference challenging. Inspired by recent advances in efficient language modeling, we propose linear recurrent units for sequential recommendation (LRURec). Similar to recurrent neural networks, LRURec offers rapid inference and can achieve incremental inference on sequential inputs. By decomposing the linear recurrence operation and designing recursive parallelization in our framework, LRURec provides the additional benefits of reduced model size and parallelizable training. Moreover, we optimize the architecture of LRURec by implementing a series of modifications to address the lack of non-linearity and improve training dynamics. To validate the effectiveness of our proposed LRURec, we conduct extensive experiments on multiple real-world datasets and compare its performance against state-of-the-art sequential recommenders. Experimental results demonstrate the effectiveness of LRURec, which consistently outperforms baselines by a significant margin. Results also highlight the efficiency of LRURec with our parallelized training paradigm and fast inference on long sequences, showing its potential to further enhance user experience in sequential recommendation.
This paper presents a classification framework based on learnable data augmentation to tackle the One-Shot Unsupervised Domain Adaptation (OS-UDA) problem. OS-UDA is the most challenging setting in Domain Adaptation, as only one single unlabeled target sample is assumed to be available for model adaptation. Driven by such single sample, our method LearnAug-UDA learns how to augment source data, making it perceptually similar to the target. As a result, a classifier trained on such augmented data will generalize well for the target domain. To achieve this, we designed an encoder-decoder architecture that exploits a perceptual loss and style transfer strategies to augment the source data. Our method achieves state-of-the-art performance on two well-known Domain Adaptation benchmarks, DomainNet and VisDA. The project code is available at //github.com/IIT-PAVIS/LearnAug-UDA
The vast majority of techniques to train fair models require access to the protected attribute (e.g., race, gender), either at train time or in production. However, in many important applications this protected attribute is largely unavailable. In this paper, we develop methods for measuring and reducing fairness violations in a setting with limited access to protected attribute labels. Specifically, we assume access to protected attribute labels on a small subset of the dataset of interest, but only probabilistic estimates of protected attribute labels (e.g., via Bayesian Improved Surname Geocoding) for the rest of the dataset. With this setting in mind, we propose a method to estimate bounds on common fairness metrics for an existing model, as well as a method for training a model to limit fairness violations by solving a constrained non-convex optimization problem. Unlike similar existing approaches, our methods take advantage of contextual information -- specifically, the relationships between a model's predictions and the probabilistic prediction of protected attributes, given the true protected attribute, and vice versa -- to provide tighter bounds on the true disparity. We provide an empirical illustration of our methods using voting data. First, we show our measurement method can bound the true disparity up to 5.5x tighter than previous methods in these applications. Then, we demonstrate that our training technique effectively reduces disparity while incurring lesser fairness-accuracy trade-offs than other fair optimization methods with limited access to protected attributes.
Classical pixel-based Visual Servoing (VS) approaches offer high accuracy but suffer from a limited convergence area due to optimization nonlinearity. Modern deep learning-based VS methods overcome traditional vision issues but lack scalability, requiring training on limited scenes. This paper proposes a hybrid VS strategy utilizing Deep Reinforcement Learning (DRL) and optimal control to enhance both convergence area and scalability. The DRL component of our approach separately handles representation and policy learning to enhance scalability, generalizability, learning efficiency and ease domain adaptation. Moreover, the optimal control part ensures high end-point accuracy. Our method showcases remarkable achievements in terms of high convergence rates and minimal end-positioning errors using a 7-DOF manipulator. Importantly, it exhibits scalability across more than 1000 distinct scenes. Furthermore, we demonstrate its capacity for generalization to previously unseen datasets. Lastly, we illustrate the real-world applicability of our approach, highlighting its adaptability through single-shot domain transfer learning in environments with noise and occlusions. Real-robot experiments can be found at \url{//sites.google.com/view/vsls}.
The interaction of fibers in a viscous (Stokes) fluid plays a crucial role in industrial and biological processes, such as sedimentation, rheology, transport, cell division, and locomotion. Numerical simulations generally rely on slender body theory (SBT), an asymptotic, nonconvergent approximation whose error blows up as fibers approach each other. Yet convergent boundary integral equation (BIE) methods which completely resolve the fiber surface have so far been impractical due to the prohibitive cost of layer-potential quadratures in such high aspect-ratio 3D geometries. We present a high-order Nystr\"om quadrature scheme with aspect-ratio independent cost, making such BIEs practical. It combines centerline panels (each with a small number of poloidal Fourier modes), toroidal Green's functions, generalized Chebyshev quadratures, HPC parallel implementation, and FMM acceleration. We also present new BIE formulations for slender bodies that lead to well conditioned linear systems upon discretization. We test Laplace and Stokes Dirichlet problems, and Stokes mobility problems, for slender rigid closed fibers with (possibly varying) circular cross-section, at separations down to $1/20$ of the slender radius, reporting convergence typically to at least 10 digits. We use this to quantify the breakdown of numerical SBT for close-to-touching rigid fibers. We also apply the methods to time-step the sedimentation of 512 loops with up to $1.65$ million unknowns at around 7 digits of accuracy.
The use of Autonomous Surface Vessels (ASVs) is growing rapidly. For safe and efficient surface auto-driving, a reliable perception system is crucial. Such systems allow the vessels to sense their surroundings and make decisions based on the information gathered. During the perception process, free space segmentation is essential to distinguish the safe mission zone and segment the operational waterways. However, ASVs face particular challenges in free space segmentation due to nearshore reflection interference, complex water textures, and random motion vibrations caused by the water surface conditions. To deal with these challenges, we propose a visual temporal fusion based free space segmentation model to utilize the previous vision information. In addition, we also introduce a new evaluation procedure and a contour position based loss calculation function, which are more suitable for surface free space segmentation tasks. The proposed model and process are tested on a continuous video segmentation dataset and achieve both high-accuracy and robust results. The dataset is also made available along with this paper.
In this paper, we introduce the Reinforced Mnemonic Reader for machine reading comprehension tasks, which enhances previous attentive readers in two aspects. First, a reattention mechanism is proposed to refine current attentions by directly accessing to past attentions that are temporally memorized in a multi-round alignment architecture, so as to avoid the problems of attention redundancy and attention deficiency. Second, a new optimization approach, called dynamic-critical reinforcement learning, is introduced to extend the standard supervised method. It always encourages to predict a more acceptable answer so as to address the convergence suppression problem occurred in traditional reinforcement learning algorithms. Extensive experiments on the Stanford Question Answering Dataset (SQuAD) show that our model achieves state-of-the-art results. Meanwhile, our model outperforms previous systems by over 6% in terms of both Exact Match and F1 metrics on two adversarial SQuAD datasets.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191