In this paper, we focus on addressing the constraints faced when applying LLMs to ASR. Recent works utilize prefixLM-type models, which directly apply speech as a prefix to LLMs for ASR. We have found that optimizing speech prefixes leads to better ASR performance and propose applying RNNT loss to perform speech prefix-tuning. This is a simple approach and does not increase the model complexity or alter the inference pipeline. We also propose language-based soft prompting to further improve with frozen LLMs. Empirical analysis on realtime testset from 10 Indic languages demonstrate that our proposed speech prefix-tuning yields improvements with both frozen and fine-tuned LLMs. Our recognition results on an average of 10 Indics show that the proposed prefix-tuning with RNNT loss results in a 12\% relative improvement in WER over the baseline with a fine-tuned LLM. Our proposed approches with the frozen LLM leads to a 31\% relative improvement over basic soft-prompting prefixLM.
相關內容
This paper explores an innovative approach to Environmental, Social, and Governance (ESG) scoring by integrating Natural Language Processing (NLP) techniques with Item Response Theory (IRT), specifically the Rasch model. The study utilizes a comprehensive dataset of news articles in Portuguese related to Petrobras, a major oil company in Brazil, collected from 2022 and 2023. The data is filtered and classified for ESG-related sentiments using advanced NLP methods. The Rasch model is then applied to evaluate the psychometric properties of these ESG measures, providing a nuanced assessment of ESG sentiment trends over time. The results demonstrate the efficacy of this methodology in offering a more precise and reliable measurement of ESG factors, highlighting significant periods and trends. This approach may enhance the robustness of ESG metrics and contribute to the broader field of sustainability and finance by offering a deeper understanding of the temporal dynamics in ESG reporting.
In this paper, the authors introduce a lightweight dataset to interpret IoT (Internet of Things) activity in preparation to create decoys by replicating known data traffic patterns. The dataset comprises different scenarios in a real network setting. This paper also surveys information related to other IoT datasets along with the characteristics that make our data valuable. Many of the datasets available are synthesized (simulated) or often address industrial applications, while the IoT dataset we present is based on likely smart home scenarios. Further, there are only a limited number of IoT datasets that contain both normal operation and attack scenarios. A discussion of the network configuration and the steps taken to prepare this dataset are presented as we prepare to create replicative patterns for decoy purposes. The dataset, which we refer to as IoT Flex Data, consists of four categories, namely, IoT benign idle, IoT benign active, IoT setup, and malicious (attack) traffic associating the IoT devices with the scenarios under consideration.
This paper proposes a Conflict-aware Resource-Efficient Decentralized Sequential planner (CREDS) for early wildfire mitigation using multiple heterogeneous Unmanned Aerial Vehicles (UAVs). Multi-UAV wildfire management scenarios are non-stationary, with spatially clustered dynamically spreading fires, potential pop-up fires, and partial observability due to limited UAV numbers and sensing range. The objective of CREDS is to detect and sequentially mitigate all growing fires as Single-UAV Tasks (SUT), minimizing biodiversity loss through rapid UAV intervention and promoting efficient resource utilization by avoiding complex multi-UAV coordination. CREDS employs a three-phased approach, beginning with fire detection using a search algorithm, followed by local trajectory generation using the auction-based Resource-Efficient Decentralized Sequential planner (REDS), incorporating the novel non-stationary cost function, the Deadline-Prioritized Mitigation Cost (DPMC). Finally, a conflict-aware consensus algorithm resolves conflicts to determine a global trajectory for spatiotemporal mitigation. The performance evaluation of the CREDS for partial and full observability conditions with both heterogeneous and homogeneous UAV teams for different fires-to-UAV ratios demonstrates a $100\%$ success rate for ratios up to $4$ and a high success rate for the critical ratio of $5$, outperforming baselines. Heterogeneous UAV teams outperform homogeneous teams in handling heterogeneous deadlines of SUT mitigation. CREDS exhibits scalability and $100\%$ convergence, demonstrating robustness against potential deadlock assignments, enhancing its success rate compared to the baseline approaches.
Diffusion models have been extensively utilized in AI-generated content (AIGC) in recent years, thanks to the superior generation capabilities. Combining with semantic communications, diffusion models are used for tasks such as denoising, data reconstruction, and content generation. However, existing diffusion-based generative models do not consider the stringent bandwidth limitation, which limits its application in wireless communication. This paper introduces a diffusion-driven semantic communication framework with advanced VAE-based compression for bandwidth-constrained generative model. Our designed architecture utilizes the diffusion model, where the signal transmission process through the wireless channel acts as the forward process in diffusion. To reduce bandwidth requirements, we incorporate a downsampling module and a paired upsampling module based on a variational auto-encoder with reparameterization at the receiver to ensure that the recovered features conform to the Gaussian distribution. Furthermore, we derive the loss function for our proposed system and evaluate its performance through comprehensive experiments. Our experimental results demonstrate significant improvements in pixel-level metrics such as peak signal to noise ratio (PSNR) and semantic metrics like learned perceptual image patch similarity (LPIPS). These enhancements are more profound regarding the compression rates and SNR compared to deep joint source-channel coding (DJSCC).
In this paper, we experimented with the SpeechT5 model pre-trained on large-scale datasets. We pre-trained the foundation model from scratch and fine-tuned it on a large-scale robust multi-speaker text-to-speech (TTS) task. We tested the model capabilities in a zero- and few-shot scenario. Based on two listening tests, we evaluated the synthetic audio quality and the similarity of how synthetic voices resemble real voices. Our results showed that the SpeechT5 model can generate a synthetic voice for any speaker using only one minute of the target speaker's data. We successfully demonstrated the high quality and similarity of our synthetic voices on publicly known Czech politicians and celebrities.
Traditional knowledge graph (KG) completion models learn embeddings to predict missing facts. Recent works attempt to complete KGs in a text-generation manner with large language models (LLMs). However, they need to ground the output of LLMs to KG entities, which inevitably brings errors. In this paper, we present a finetuning framework, DIFT, aiming to unleash the KG completion ability of LLMs and avoid grounding errors. Given an incomplete fact, DIFT employs a lightweight model to obtain candidate entities and finetunes an LLM with discrimination instructions to select the correct one from the given candidates. To improve performance while reducing instruction data, DIFT uses a truncated sampling method to select useful facts for finetuning and injects KG embeddings into the LLM. Extensive experiments on benchmark datasets demonstrate the effectiveness of our proposed framework.
This paper focuses on hyperparameter optimization for autonomous driving strategies based on Reinforcement Learning. We provide a detailed description of training the RL agent in a simulation environment. Subsequently, we employ Efficient Global Optimization algorithm that uses Gaussian Process fitting for hyperparameter optimization in RL. Before this optimization phase, Gaussian process interpolation is applied to fit the surrogate model, for which the hyperparameter set is generated using Latin hypercube sampling. To accelerate the evaluation, parallelization techniques are employed. Following the hyperparameter optimization procedure, a set of hyperparameters is identified, resulting in a noteworthy enhancement in overall driving performance. There is a substantial increase of 4\% when compared to existing manually tuned parameters and the hyperparameters discovered during the initialization process using Latin hypercube sampling. After the optimization, we analyze the obtained results thoroughly and conduct a sensitivity analysis to assess the robustness and generalization capabilities of the learned autonomous driving strategies. The findings from this study contribute to the advancement of Gaussian process based Bayesian optimization to optimize the hyperparameters for autonomous driving in RL, providing valuable insights for the development of efficient and reliable autonomous driving systems.
To fill the gap of traditional GS compression method, in this paper, we first propose a simple and effective GS data compression anchor called Graph-based GS Compression (GGSC). GGSC is inspired by graph signal processing theory and uses two branches to compress the primitive center and attributes. We split the whole GS sample via KDTree and clip the high-frequency components after the graph Fourier transform. Followed by quantization, G-PCC and adaptive arithmetic coding are used to compress the primitive center and attribute residual matrix to generate the bitrate file. GGSS is the first work to explore traditional GS compression, with advantages that can reveal the GS distortion characteristics corresponding to typical compression operation, such as high-frequency clipping and quantization. Second, based on GGSC, we create a GS Quality Assessment dataset (GSQA) with 120 samples. A subjective experiment is conducted in a laboratory environment to collect subjective scores after rendering GS into Processed Video Sequences (PVS). We analyze the characteristics of different GS distortions based on Mean Opinion Scores (MOS), demonstrating the sensitivity of different attributes distortion to visual quality. The GGSC code and the dataset, including GS samples, MOS, and PVS, are made publicly available at //github.com/Qi-Yangsjtu/GGSC.
In this paper, we proposed to apply meta learning approach for low-resource automatic speech recognition (ASR). We formulated ASR for different languages as different tasks, and meta-learned the initialization parameters from many pretraining languages to achieve fast adaptation on unseen target language, via recently proposed model-agnostic meta learning algorithm (MAML). We evaluated the proposed approach using six languages as pretraining tasks and four languages as target tasks. Preliminary results showed that the proposed method, MetaASR, significantly outperforms the state-of-the-art multitask pretraining approach on all target languages with different combinations of pretraining languages. In addition, since MAML's model-agnostic property, this paper also opens new research direction of applying meta learning to more speech-related applications.
In this paper, we propose the joint learning attention and recurrent neural network (RNN) models for multi-label classification. While approaches based on the use of either model exist (e.g., for the task of image captioning), training such existing network architectures typically require pre-defined label sequences. For multi-label classification, it would be desirable to have a robust inference process, so that the prediction error would not propagate and thus affect the performance. Our proposed model uniquely integrates attention and Long Short Term Memory (LSTM) models, which not only addresses the above problem but also allows one to identify visual objects of interests with varying sizes without the prior knowledge of particular label ordering. More importantly, label co-occurrence information can be jointly exploited by our LSTM model. Finally, by advancing the technique of beam search, prediction of multiple labels can be efficiently achieved by our proposed network model.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191