South America
LLaMA-Omni: Seamless Speech Interaction with Large Language Models
Fang, Qingkai, Guo, Shoutao, Zhou, Yan, Ma, Zhengrui, Zhang, Shaolei, Feng, Yang
Models like GPT-4o enable real-time interaction with large language models (LLMs) through speech, significantly enhancing user experience compared to traditional text-based interaction. However, there is still a lack of exploration on how to build speech interaction models based on open-source LLMs. To address this, we propose LLaMA-Omni, a novel model architecture designed for low-latency and high-quality speech interaction with LLMs. It eliminates the need for speech transcription, and can simultaneously generate text and speech responses directly from speech instructions with extremely low latency. We build our model based on the latest Llama-3.1-8B-Instruct To align the model with speech interaction scenarios, we construct a dataset named InstructS2S-200K, which includes 200K speech instructions and corresponding speech responses. Experimental results show that compared to previous speech-language models, LLaMA-Omni provides better responses in both content and style, with a response latency as low as 226ms. Additionally, training LLaMA-Omni takes less than 3 days on just 4 GPUs, paving the way for the efficient development of speech-language models in the future. Large language models (LLMs), represented by ChatGPT (OpenAI, 2022), have become powerful general-purpose task solvers, capable of assisting people in daily life through conversational interactions. However, most LLMs currently only support text-based interactions, which limits their application in scenarios where text input and output are not ideal. Recently, the emergence of GPT-4o (OpenAI, 2024) has made it possible to interact with LLMs through speech, responding to user's instruction with extremely low latency and significantly enhancing the user experience.
Retrieval Or Holistic Understanding? Dolce: Differentiate Our Long Context Evaluation Tasks
We argue that there are two major distinct capabilities in long context understanding: retrieval and holistic understanding. Understanding and further improving LLMs' long context capabilities would not be possible without knowing the tasks' focus categories. We aim to automatically identify retrieval focused and holistic understanding focused problems from suites of benchmarks and quantitatively measure the difficulty within each focus. In this paper, we present the Dolce framework, which parameterizes each problem by $\lambda$ (complexity) and $k$ (redundancy) and assigns to one of five predefined focus categories. We propose to sample short contexts from the full context and estimate the probability an LLM solves the problem using the sampled spans. To find the $\lambda$ and $k$ for each problem, we further propose a mixture model of a non-parametric background noise component and a parametric/non-parametric hybrid oracle component, where we derive the probability functions parameterized by $\lambda$ and $k$ for both the correct-or-wrong (COW) scenario and the partial-point-in-grading (PIG) scenario. Our proposed methods can identify 0% to 67% of the problems are retrieval focused and 0% to 90% of the problems are holistic understanding focused across 44 existing long context evaluation tasks.
Length Desensitization in Directed Preference Optimization
Liu, Wei, Bai, Yang, Han, Chengcheng, Weng, Rongxiang, Xu, Jun, Cao, Xuezhi, Wang, Jingang, Cai, Xunliang
Direct Preference Optimization (DPO) is widely utilized in the Reinforcement Learning from Human Feedback (RLHF) phase to align Large Language Models (LLMs) with human preferences, thereby enhancing both their harmlessness and efficacy. However, it has been observed that DPO tends to over-optimize for verbosity, which can detrimentally affect both performance and user experience. In this paper, we conduct an in-depth theoretical analysis of DPO's optimization objective and reveal a strong correlation between its implicit reward and data length. To address this issue, we propose a length-desensitization improvement method for DPO, termed LD-DPO. The proposed method aims to desensitize DPO to data length by decoupling explicit length preference, which is relatively insignificant, from the other implicit preferences, thereby enabling more effective learning of the intrinsic preferences. We utilized two settings (Base and Instruct) of Llama2-13B, Llama3-8B, and Qwen2-7B for experimental validation on various benchmarks including MT-Bench and AlpacaEval 2. The experimental results indicate that LD-DPO consistently outperforms DPO and other baseline methods, achieving more concise responses with a 10-40% reduction in length compared to DPO. We conducted in-depth experimental analyses to demonstrate that LD-DPO can indeed achieve length desensitization and align the model more closely with human-real preferences. "Brevity is the Soul of Wit." Human preference alignment is crucial to enable large language models (LLMs) to be helpful, honest, and harmless. Among the various methods to achieve effective alignment (Dai et al., 2024; Yuan et al., 2024a), Directed Preference Optimization (DPO) has emerged as a promising technique (Rafailov et al., 2024), giving rise to numerous derivative algorithms (Hong et al., 2024; Chen et al., 2024b; Ethayarajh et al., 2024). DPO eliminates the reliance on online Reward Models (RMs) by reparameterizing the reward function in Reinforcement Learning from Human Feedback (RLHF), thereby implementing a simple and stable offline preference learning paradigm. Among the dimensions of human language preferences, detailedness is one of the most straightforward categories that current alignment algorithms can effortlessly capture, as longer texts tend to be richer in content. However, it has been demonstrated that DPO is susceptible to an over-optimization issue in this particular preference dimension (Xu et al., 2024). As shown in Fig.1, this overemphasis results in models that produce excessively verbose responses, which can compromise their instruction-following and reasoning capabilities (Ding et al., 2023; Yuan et al., 2024b).
Quantifying and Enabling the Interpretability of CLIP-like Models
Madasu, Avinash, Gandelsman, Yossi, Lal, Vasudev, Howard, Phillip
CLIP is one of the most popular foundational models and is heavily used for many vision-language tasks. However, little is known about the inner workings of CLIP. To bridge this gap we propose a study to quantify the interpretability in CLIP like models. We conduct this study on six different CLIP models from OpenAI and OpenCLIP which vary by size, type of pre-training data and patch size. Our approach begins with using the TEXTSPAN algorithm and in-context learning to break down individual attention heads into specific properties. We then evaluate how easily these heads can be interpreted using new metrics which measure property consistency within heads and property disentanglement across heads. Our findings reveal that larger CLIP models are generally more interpretable than their smaller counterparts. To further assist users in understanding the inner workings of CLIP models, we introduce CLIP-InterpreT, a tool designed for interpretability analysis. CLIP-InterpreT offers five types of analyses: property-based nearest neighbor search, per-head topic segmentation, contrastive segmentation, per-head nearest neighbors of an image, and per-head nearest neighbors of text.
VITA: Towards Open-Source Interactive Omni Multimodal LLM
Fu, Chaoyou, Lin, Haojia, Long, Zuwei, Shen, Yunhang, Zhao, Meng, Zhang, Yifan, Dong, Shaoqi, Wang, Xiong, Yin, Di, Ma, Long, Zheng, Xiawu, He, Ran, Ji, Rongrong, Wu, Yunsheng, Shan, Caifeng, Sun, Xing
The remarkable multimodal capabilities and interactive experience of GPT-4o underscore their necessity in practical applications, yet open-source models rarely excel in both areas. In this paper, we introduce VITA, the first-ever open-source Multimodal Large Language Model (MLLM) adept at simultaneous processing and analysis of Video, Image, Text, and Audio modalities, and meanwhile has an advanced multimodal interactive experience. Starting from Mixtral 8x7B as a language foundation, we expand its Chinese vocabulary followed by bilingual instruction tuning. We further endow the language model with visual and audio capabilities through two-stage multi-task learning of multimodal alignment and instruction tuning. VITA demonstrates robust foundational capabilities of multilingual, vision, and audio understanding, as evidenced by its strong performance across a range of both unimodal and multimodal benchmarks. Beyond foundational capabilities, we have made considerable progress in enhancing the natural multimodal human-computer interaction experience. VITA is the first step for the open-source community to explore the seamless integration of multimodal understanding and interaction. While there is still lots of work to be done on VITA to get close to close-source counterparts, we hope that its role as a pioneer can serve as a cornerstone for subsequent research. Project Page: https://vita-home.github.io.
CerviXpert: A Multi-Structural Convolutional Neural Network for Predicting Cervix Type and Cervical Cell Abnormalities
Akash, Rashik Shahriar, Islam, Radiful, Badhon, S. M. Saiful Islam, Hossain, K. S. M. Tozammel
Cervical cancer affects millions of women worldwide and has a significantly higher survival rate when diagnosed early. Pap smears and cervical biopsies are vital screening tools for detecting such cancer. However, the success of these screening processes depends on the skills of cytologists. A recent trend in diagnostic cytology is to apply machine-learning-based models to classify cancer using cell images. These automated models have been shown to perform just as well as, or even better than, expert cytologists. Some notable methods for classifying cervix cancers include ResNet50, VGG16, MobileNetV2, and InceptionV3, based on deep convolutional neural networks (CNN). However, these methods are computationally expensive. We present CerviXpert, a multi-structural Convolutional Neural Network, to identify cervix cancer. We perform extensive experiments on a publicly available dataset, SiPaKMeD, to show the efficacy of our method. CerviXpert presents a promising solution for efficient cervical cancer screening and diagnosis by striking a balance between accuracy and practical feasibility.
MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark
Yue, Xiang, Zheng, Tianyu, Ni, Yuansheng, Wang, Yubo, Zhang, Kai, Tong, Shengbang, Sun, Yuxuan, Yu, Botao, Zhang, Ge, Sun, Huan, Su, Yu, Chen, Wenhu, Neubig, Graham
This paper introduces MMMU-Pro, a robust version of the Massive Multi-discipline Multimodal Understanding and Reasoning (MMMU) benchmark. MMMU-Pro rigorously assesses multimodal models' true understanding and reasoning capabilities through a three-step process based on MMMU: (1) filtering out questions answerable by text-only models, (2) augmenting candidate options, and (3) introducing a vision-only input setting where questions are embedded within images. This setting challenges AI to truly "see" and "read" simultaneously, testing a fundamental human cognitive skill of seamlessly integrating visual and textual information. Results show that model performance is substantially lower on MMMU-Pro than on MMMU, ranging from 16.8% to 26.9% across models. We explore the impact of OCR prompts and Chain of Thought (CoT) reasoning, finding that OCR prompts have minimal effect while CoT generally improves performance. MMMU-Pro provides a more rigorous evaluation tool, closely mimicking real-world scenarios and offering valuable directions for future research in multimodal AI.
Check-Eval: A Checklist-based Approach for Evaluating Text Quality
Pereira, Jayr, Assumpcao, Andre, Lotufo, Roberto
Evaluating the quality of text generated by large language models (LLMs) remains a significant challenge. Traditional metrics often fail to align well with human judgments, particularly in tasks requiring creativity and nuance. In this paper, we propose \textsc{Check-Eval}, a novel evaluation framework leveraging LLMs to assess the quality of generated text through a checklist-based approach. \textsc{Check-Eval} can be employed as both a reference-free and reference-dependent evaluation method, providing a structured and interpretable assessment of text quality. The framework consists of two main stages: checklist generation and checklist evaluation. We validate \textsc{Check-Eval} on two benchmark datasets: Portuguese Legal Semantic Textual Similarity and \textsc{SummEval}. Our results demonstrate that \textsc{Check-Eval} achieves higher correlations with human judgments compared to existing metrics, such as \textsc{G-Eval} and \textsc{GPTScore}, underscoring its potential as a more reliable and effective evaluation framework for natural language generation tasks. The code for our experiments is available at \url{https://anonymous.4open.science/r/check-eval-0DB4}
A Large Dataset of Spontaneous Speech with the Accent Spoken in S\~ao Paulo for Automatic Speech Recognition Evaluation
Lima, Rodrigo, Leal, Sidney Evaldo, Junior, Arnaldo Candido, Aluísio, Sandra Maria
We present a freely available spontaneous speech corpus for the Brazilian Portuguese language and report preliminary automatic speech recognition (ASR) results, using both the Wav2Vec2-XLSR-53 and Distil-Whisper models fine-tuned and trained on our corpus. The NURC-SP Audio Corpus comprises 401 different speakers (204 females, 197 males) with a total of 239.30 hours of transcribed audio recordings. To the best of our knowledge, this is the first large Paulistano accented spontaneous speech corpus dedicated to the ASR task in Portuguese. We first present the design and development procedures of the NURC-SP Audio Corpus, and then describe four ASR experiments in detail. The experiments demonstrated promising results for the applicability of the corpus for ASR. Specifically, we fine-tuned two versions of Wav2Vec2-XLSR-53 model, trained a Distil-Whisper model using our dataset with labels determined by Whisper Large-V3 model, and fine-tuned this Distil-Whisper model with our corpus. Our best results were the Distil-Whisper fine-tuned over NURC-SP Audio Corpus with a WER of 24.22% followed by a fine-tuned versions of Wav2Vec2-XLSR-53 model with a WER of 33.73%, that is almost 10% point worse than Distil-Whisper's. To enable experiment reproducibility, we share the NURC-SP Audio Corpus dataset, pre-trained models, and training recipes in Hugging-Face and Github repositories.
Damage detection in an uncertain nonlinear beam based on stochastic Volterra series
Villani, Luis Gustavo Giacon, da Silva, Samuel, Cunha, Americo Jr
The damage detection problem in mechanical systems, using vibration measurements, is commonly called Structural Health Monitoring (SHM). Many tools are able to detect damages by changes in the vibration pattern, mainly, when damages induce nonlinear behavior. However, a more difficult problem is to detect structural variation associated with damage, when the mechanical system has nonlinear behavior even in the reference condition. In these cases, more sophisticated methods are required to detect if the changes in the response are based on some structural variation or changes in the vibration regime, because both can generate nonlinearities. Among the many ways to solve this problem, the use of the Volterra series has several favorable points, because they are a generalization of the linear convolution, allowing the separation of linear and nonlinear contributions by input filtering through the Volterra kernels. On the other hand, the presence of uncertainties in mechanical systems, due to noise, geometric imperfections, manufacturing irregularities, environmental conditions, and others, can also change the responses, becoming more difficult the damage detection procedure. An approach based on a stochastic version of Volterra series is proposed to be used in the detection of a breathing crack in a beam vibrating in a nonlinear regime of motion, even in reference condition (without crack). The system uncertainties are simulated by the variation imposed in the linear stiffness and damping coefficient. The results show, that the nonlinear analysis done, considering the high order Volterra kernels, allows the approach to detect the crack with a small propagation and probability confidence, even in the presence of uncertainties.