Conformer-based end-to-end models have become ubiquitous these days and are commonly used in both streaming and non-streaming automatic speech recognition (ASR). Techniques like dual-mode and dynamic chunk training helped unify streaming and non-streaming systems. However, there remains a performance gap between streaming with a full and limited past context. To address this issue, we propose the integration of a novel dynamic contextual carry-over mechanism in a state-of-the-art (SOTA) unified ASR system. Our proposed dynamic context Conformer (DCTX-Conformer) utilizes a non-overlapping contextual carry-over mechanism that takes into account both the left context of a chunk and one or more preceding context embeddings. We outperform the SOTA by a relative 25.0% word error rate, with a negligible latency impact due to the additional context embeddings.
相關內容
Neural speech codec has recently gained widespread attention in generative speech modeling domains, like voice conversion, text-to-speech synthesis, etc. However, ensuring high-fidelity audio reconstruction of speech codecs under low bitrate remains an open and challenging issue. In this paper, we propose PromptCodec, a novel end-to-end neural speech codec using feature-aware prompt encoders based on disentangled representation learning. By incorporating prompt encoders to capture representations of additional input prompts, PromptCodec can distribute the speech information requiring processing and enhance its representation capabilities. Moreover, a simple yet effective adaptive feature weighted fusion approach is introduced to integrate features of different encoders. Meanwhile, we propose a novel disentangled representation learning strategy based on structure similarity index measure to optimize PromptCodec's encoders to ensure their efficiency, thereby further improving the performance of PromptCodec. Experiments on LibriTTS demonstrate that our proposed PromptCodec consistently outperforms state-of-the-art neural speech codec models under all different bitrate conditions while achieving superior performance with low bitrates.
Significant strides have been made in advancing streamflow predictions, notably with the introduction of cutting-edge machine-learning models. Predominantly, Long Short-Term Memories (LSTMs) and Convolution Neural Networks (CNNs) have been widely employed in this domain. While LSTMs are applicable in both rainfall-runoff and time series settings, CNN-LSTMs have primarily been utilized in rainfall-runoff scenarios. In this study, we extend the application of CNN-LSTMs to time series settings, leveraging lagged streamflow data in conjunction with precipitation and temperature data to predict streamflow. Our results show a substantial improvement in predictive performance in 21 out of 32 HUC8 basins in Nebraska, showcasing noteworthy increases in the Kling-Gupta Efficiency (KGE) values. These results highlight the effectiveness of CNN-LSTMs in time series settings, particularly for spatiotemporal hydrological modeling, for more accurate and robust streamflow predictions.
Aligning a user query and video clips in cross-modal latent space and that with semantic concepts are two mainstream approaches for ad-hoc video search (AVS). However, the effectiveness of existing approaches is bottlenecked by the small sizes of available video-text datasets and the low quality of concept banks, which results in the failures of unseen queries and the out-of-vocabulary problem. This paper addresses these two problems by constructing a new dataset and developing a multi-word concept bank. Specifically, capitalizing on a generative model, we construct a new dataset consisting of 7 million generated text and video pairs for pre-training. To tackle the out-of-vocabulary problem, we develop a multi-word concept bank based on syntax analysis to enhance the capability of a state-of-the-art interpretable AVS method in modeling relationships between query words. We also study the impact of current advanced features on the method. Experimental results show that the integration of the above-proposed elements doubles the R@1 performance of the AVS method on the MSRVTT dataset and improves the xinfAP on the TRECVid AVS query sets for 2016-2023 (eight years) by a margin from 2% to 77%, with an average about 20%.
The embedding-based retrieval (EBR) approach is widely used in mainstream search engine retrieval systems and is crucial in recent retrieval-augmented methods for eliminating LLM illusions. However, existing EBR models often face the "semantic drift" problem and insufficient focus on key information, leading to a low adoption rate of retrieval results in subsequent steps. This issue is especially noticeable in real-time search scenarios, where the various expressions of popular events on the Internet make real-time retrieval heavily reliant on crucial event information. To tackle this problem, this paper proposes a novel approach called EER, which enhances real-time retrieval performance by improving the dual-encoder model of traditional EBR. We incorporate contrastive learning to accompany pairwise learning for encoder optimization. Furthermore, to strengthen the focus on critical event information in events, we include a decoder module after the document encoder, introduce a generative event triplet extraction scheme based on prompt-tuning, and correlate the events with query encoder optimization through comparative learning. This decoder module can be removed during inference. Extensive experiments demonstrate that EER can significantly improve the real-time search retrieval performance. We believe that this approach will provide new perspectives in the field of information retrieval. The codes and dataset are available at //github.com/open-event-hub/Event-enhanced_Retrieval .
Cross-lingual image captioning is a challenging task that requires addressing both cross-lingual and cross-modal obstacles in multimedia analysis. The crucial issue in this task is to model the global and the local matching between the image and different languages. Existing cross-modal embedding methods based on the transformer architecture oversee the local matching between the image region and monolingual words, especially when dealing with diverse languages. To overcome these limitations, we propose an Embedded Heterogeneous Attention Transformer (EHAT) to establish cross-domain relationships and local correspondences between images and different languages by using a heterogeneous network. EHAT comprises Masked Heterogeneous Cross-attention (MHCA), Heterogeneous Attention Reasoning Network (HARN), and Heterogeneous Co-attention (HCA). The HARN serves as the core network and it captures cross-domain relationships by leveraging visual bounding box representation features to connect word features from two languages and to learn heterogeneous maps. MHCA and HCA facilitate cross-domain integration in the encoder through specialized heterogeneous attention mechanisms, enabling a single model to generate captions in two languages. We evaluate our approach on the MSCOCO dataset to generate captions in English and Chinese, two languages that exhibit significant differences in their language families. The experimental results demonstrate the superior performance of our method compared to existing advanced monolingual methods. Our proposed EHAT framework effectively addresses the challenges of cross-lingual image captioning, paving the way for improved multilingual image analysis and understanding.
In the field of multi-object tracking (MOT), recent Transformer based end-to-end models like MOTR have demonstrated exceptional performance on datasets such as DanceTracker. However, the computational demands of these models present challenges in training and deployment. Drawing inspiration from successful models like GPT, we present MO-YOLO, an efficient and computationally frugal end-to-end MOT model. MO-YOLO integrates principles from You Only Look Once (YOLO) and RT-DETR, adopting a decoder-only approach. By leveraging the decoder from RT-DETR and architectural components from YOLOv8, MO-YOLO achieves high speed, shorter training times, and proficient MOT performance. On the Dancetrack, MO-YOLO not only matches MOTR's performance but also surpasses it, achieving over twice the frames per second (MOTR 9.5 FPS, MO-YOLO 19.6 FPS). Furthermore, MO-YOLO demonstrates significantly reduced training times and lower hardware requirements compared to MOTR. This research introduces a promising paradigm for efficient end-to-end MOT, emphasizing enhanced performance and resource efficiency.
Recent advances in generative diffusion models have enabled the previously unfeasible capability of generating 3D assets from a single input image or a text prompt. In this work, we aim to enhance the quality and functionality of these models for the task of creating controllable, photorealistic human avatars. We achieve this by integrating a 3D morphable model into the state-of-the-art multi-view-consistent diffusion approach. We demonstrate that accurate conditioning of a generative pipeline on the articulated 3D model enhances the baseline model performance on the task of novel view synthesis from a single image. More importantly, this integration facilitates a seamless and accurate incorporation of facial expression and body pose control into the generation process. To the best of our knowledge, our proposed framework is the first diffusion model to enable the creation of fully 3D-consistent, animatable, and photorealistic human avatars from a single image of an unseen subject; extensive quantitative and qualitative evaluations demonstrate the advantages of our approach over existing state-of-the-art avatar creation models on both novel view and novel expression synthesis tasks. The code for our project is publicly available.
FlashAttention (Dao, 2023) effectively reduces the quadratic peak memory usage to linear in training transformer-based large language models (LLMs) on a single GPU. In this paper, we introduce DISTFLASHATTN, a distributed memory-efficient attention mechanism optimized for long-context LLMs training. We propose three key techniques: token-level workload balancing, overlapping key-value communication, and a rematerialization-aware gradient checkpointing algorithm. We evaluate DISTFLASHATTN on Llama-7B and variants with sequence lengths from 32K to 512K. DISTFLASHATTN achieves 8x longer sequences, 4.45 - 5.64x speedup compared to Ring Self-Attention, 2 - 8x longer sequences, 1.24 - 2.01x speedup compared to Megatron-LM with FlashAttention. It achieves 1.67x and 1.26 - 1.88x speedup compared to recent Ring Attention and DeepSpeed-Ulysses. Code is available at //github.com/RulinShao/LightSeq.
Contrastive learning allows us to flexibly define powerful losses by contrasting positive pairs from sets of negative samples. Recently, the principle has also been used to learn cross-modal embeddings for video and text, yet without exploiting its full potential. In particular, previous losses do not take the intra-modality similarities into account, which leads to inefficient embeddings, as the same content is mapped to multiple points in the embedding space. With CrossCLR, we present a contrastive loss that fixes this issue. Moreover, we define sets of highly related samples in terms of their input embeddings and exclude them from the negative samples to avoid issues with false negatives. We show that these principles consistently improve the quality of the learned embeddings. The joint embeddings learned with CrossCLR extend the state of the art in video-text retrieval on Youcook2 and LSMDC datasets and in video captioning on Youcook2 dataset by a large margin. We also demonstrate the generality of the concept by learning improved joint embeddings for other pairs of modalities.
Self-supervised learning methods are gaining increasing traction in computer vision due to their recent success in reducing the gap with supervised learning. In natural language processing (NLP) self-supervised learning and transformers are already the methods of choice. The recent literature suggests that the transformers are becoming increasingly popular also in computer vision. So far, the vision transformers have been shown to work well when pretrained either using a large scale supervised data or with some kind of co-supervision, e.g. in terms of teacher network. These supervised pretrained vision transformers achieve very good results in downstream tasks with minimal changes. In this work we investigate the merits of self-supervised learning for pretraining image/vision transformers and then using them for downstream classification tasks. We propose Self-supervised vIsion Transformers (SiT) and discuss several self-supervised training mechanisms to obtain a pretext model. The architectural flexibility of SiT allows us to use it as an autoencoder and work with multiple self-supervised tasks seamlessly. We show that a pretrained SiT can be finetuned for a downstream classification task on small scale datasets, consisting of a few thousand images rather than several millions. The proposed approach is evaluated on standard datasets using common protocols. The results demonstrate the strength of the transformers and their suitability for self-supervised learning. We outperformed existing self-supervised learning methods by large margin. We also observed that SiT is good for few shot learning and also showed that it is learning useful representation by simply training a linear classifier on top of the learned features from SiT. Pretraining, finetuning, and evaluation codes will be available under: //github.com/Sara-Ahmed/SiT.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191