GPT-4o System Card

GPT-4o is an autoregressive omni model that accepts as input any combination of text, audio, image, and video, and generates any combination of text, audio, and image outputs. It's trained end-to-end across text, vision, and audio, meaning all inputs and outputs are processed by the same neural network. GPT-4o can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time in conversation. It matches GPT-4 Turbo performance on text in English and code, with significant improvement on text in non-English languages, while also being much faster and 50\% cheaper in the API. GPT-4o is especially better at vision and audio understanding compared to existing models. In line with our commitment to building AI safely and consistent with our voluntary commitments to the White House, we are sharing the GPT-4o System Card, which includes our Preparedness Framework evaluations. In this System Card, we provide a detailed look at GPT-4o's capabilities, limitations, and safety evaluations across multiple categories, focusing on speech-to-speech while also evaluating text and image capabilities, and measures we've implemented to ensure the model is safe and aligned. We also include third-party assessments on dangerous capabilities, as well as discussion of potential societal impacts of GPT-4o's text and vision capabilities.

Simplifying, Stabilizing and Scaling Continuous-Time Consistency Models

Consistency models (CMs) are a powerful class of diffusion-based generative models optimized for fast sampling. Most existing CMs are trained using discretized timesteps, which introduce additional hyperparameters and are prone to discretization errors. While continuous-time formulations can mitigate these issues, their success has been limited by training instability. To address this, we propose a simplified theoretical framework that unifies previous parameterizations of diffusion models and CMs, identifying the root causes of instability. Based on this analysis, we introduce key improvements in diffusion process parameterization, network architecture, and training objectives. These changes enable us to train continuous-time CMs at an unprecedented scale, reaching 1.5B parameters on ImageNet 512x512. Our proposed training algorithm, using only two sampling steps, achieves FID scores of 2.06 on CIFAR-10, 1.48 on ImageNet 64x64, and 1.88 on ImageNet 512x512, narrowing the gap in FID scores with the best existing diffusion models to within 10%.

Towards Realistic Emotional Voice Conversion using Controllable Emotional Intensity

Realistic emotional voice conversion (EVC) aims to enhance emotional diversity of converted audios, making the synthesized voices more authentic and natural. To this end, we propose Emotional Intensity-aware Network (EINet), dynamically adjusting intonation and rhythm by incorporating controllable emotional intensity. To better capture nuances in emotional intensity, we go beyond mere distance measurements among acoustic features. Instead, an emotion evaluator is utilized to precisely quantify speaker's emotional state. By employing an intensity mapper, intensity pseudo-labels are obtained to bridge the gap between emotional speech intensity modeling and run-time conversion. To ensure high speech quality while retaining controllability, an emotion renderer is used for combining linguistic features smoothly with manipulated emotional features at frame level. Furthermore, we employ a duration predictor to facilitate adaptive prediction of rhythm changes condition on specifying intensity value. Experimental results show EINet's superior performance in naturalness and diversity of emotional expression compared to state-of-the-art EVC methods.

Temporal Label Hierachical Network for Compound Emotion Recognition

The emotion recognition has attracted more attention in recent decades. Although significant progress has been made in the recognition technology of the seven basic emotions, existing methods are still hard to tackle compound emotion recognition that occurred commonly in practical application. This article introduces our achievements in the 7th Field Emotion Behavior Analysis (ABAW) competition. In the competition, we selected pre trained ResNet18 and Transformer, which have been widely validated, as the basic network framework. Considering the continuity of emotions over time, we propose a time pyramid structure network for frame level emotion prediction. Furthermore. At the same time, in order to address the lack of data in composite emotion recognition, we utilized fine-grained labels from the DFEW database to construct training data for emotion categories in competitions. Taking into account the characteristics of valence arousal of various complex emotions, we constructed a classification framework from coarse to fine in the label space.

Energy Model-based Accurate Shapley Value Estimation for Interpretable Deep Learning Predictive Modelling

As a favorable tool for explainable artificial intelligence (XAI), Shapley value has been widely used to interpret deep learning based predictive models. However, accurate and efficient estimation of Shapley value is a difficult task since the computation load grows exponentially with the increase of input features. Most existing accelerated Shapley value estimation methods have to compromise on estimation accuracy with efficiency. In this article, we present EmSHAP(Energy model-based Shapley value estimation), which can effectively approximate the expectation of Shapley contribution function/deep learning model under arbitrary subset of features given the rest. In order to determine the proposal conditional distribution in the energy model, a gated recurrent unit(GRU) is introduced by mapping the input features onto a hidden space, so that the impact of input feature orderings can be eliminated. In addition, a dynamic masking scheme is proposed to improve the generalization ability. It is proved in Theorems 1, 2 and 3 that EmSHAP achieves tighter error bound than state-of-the-art methods like KernelSHAP and VAEAC, leading to higher estimation accuracy. Finally, case studies on a medical application and an industrial application show that the proposed Shapley value-based explainable framework exhibits enhanced estimation accuracy without compromise on efficiency.

PAVITS: Exploring Prosody-aware VITS for End-to-End Emotional Voice Conversion

In this paper, we propose Prosody-aware VITS (PAVITS) for emotional voice conversion (EVC), aiming to achieve two major objectives of EVC: high content naturalness and high emotional naturalness, which are crucial for meeting the demands of human perception. To improve the content naturalness of converted audio, we have developed an end-to-end EVC architecture inspired by the high audio quality of VITS. By seamlessly integrating an acoustic converter and vocoder, we effectively address the common issue of mismatch between emotional prosody training and run-time conversion that is prevalent in existing EVC models. To further enhance the emotional naturalness, we introduce an emotion descriptor to model the subtle prosody variations of different speech emotions. Additionally, we propose a prosody predictor, which predicts prosody features from text based on the provided emotion label. Notably, we introduce a prosody alignment loss to establish a connection between latent prosody features from two distinct modalities, ensuring effective training. Experimental results show that the performance of PAVITS is superior to the state-of-the-art EVC methods. Speech Samples are available at https://jeremychee4.github.io/pavits4EVC/ .

Towards Efficient and Exact Optimization of Language Model Alignment

The alignment of language models with human preferences is vital for their application in real-world tasks. The problem is formulated as optimizing the model's policy to maximize the expected reward that reflects human preferences with minimal deviation from the initial policy. While considered as a straightforward solution, reinforcement learning (RL) suffers from high variance in policy updates, which impedes efficient policy improvement. Recently, direct preference optimization (DPO) was proposed to directly optimize the policy from preference data. Though simple to implement, DPO is derived based on the optimal policy that is not assured to be achieved in practice, which undermines its convergence to the intended solution. In this paper, we propose efficient exact optimization (EXO) of the alignment objective. We prove that EXO is guaranteed to optimize in the same direction as the RL algorithms asymptotically for arbitary parametrization of the policy, while enables efficient optimization by circumventing the complexities associated with RL algorithms. We compare our method to DPO with both theoretical and empirical analyses, and further demonstrate the advantages of our method over existing approaches on realistic human preference data.

Anything in Any Scene: Photorealistic Video Object Insertion

Realistic video simulation has shown significant potential across diverse applications, from virtual reality to film production. This is particularly true for scenarios where capturing videos in real-world settings is either impractical or expensive. Existing approaches in video simulation often fail to accurately model the lighting environment, represent the object geometry, or achieve high levels of photorealism. In this paper, we propose Anything in Any Scene, a novel and generic framework for realistic video simulation that seamlessly inserts any object into an existing dynamic video with a strong emphasis on physical realism. Our proposed general framework encompasses three key processes: 1) integrating a realistic object into a given scene video with proper placement to ensure geometric realism; 2) estimating the sky and environmental lighting distribution and simulating realistic shadows to enhance the light realism; 3) employing a style transfer network that refines the final video output to maximize photorealism. We experimentally demonstrate that Anything in Any Scene framework produces simulated videos of great geometric realism, lighting realism, and photorealism. By significantly mitigating the challenges associated with video data generation, our framework offers an efficient and cost-effective solution for acquiring high-quality videos. Furthermore, its applications extend well beyond video data augmentation, showing promising potential in virtual reality, video editing, and various other video-centric applications. Please check our project website https://anythinginanyscene.github.io for access to our project code and more high-resolution video results.

Emotion-Aware Contrastive Adaptation Network for Source-Free Cross-Corpus Speech Emotion Recognition

Cross-corpus speech emotion recognition (SER) aims to transfer emotional knowledge from a labeled source corpus to an unlabeled corpus. However, prior methods require access to source data during adaptation, which is unattainable in real-life scenarios due to data privacy protection concerns. This paper tackles a more practical task, namely source-free cross-corpus SER, where a pre-trained source model is adapted to the target domain without access to source data. To address the problem, we propose a novel method called emotion-aware contrastive adaptation network (ECAN). The core idea is to capture local neighborhood information between samples while considering the global class-level adaptation. Specifically, we propose a nearest neighbor contrastive learning to promote local emotion consistency among features of highly similar samples. Furthermore, relying solely on nearest neighborhoods may lead to ambiguous boundaries between clusters. Thus, we incorporate supervised contrastive learning to encourage greater separation between clusters representing different emotions, thereby facilitating improved class-level adaptation. Extensive experiments indicate that our proposed ECAN significantly outperforms state-of-the-art methods under the source-free cross-corpus SER setting on several speech emotion corpora.

Speech Swin-Transformer: Exploring a Hierarchical Transformer with Shifted Windows for Speech Emotion Recognition

Swin-Transformer has demonstrated remarkable success in computer vision by leveraging its hierarchical feature representation based on Transformer. In speech signals, emotional information is distributed across different scales of speech features, e.\,g., word, phrase, and utterance. Drawing above inspiration, this paper presents a hierarchical speech Transformer with shifted windows to aggregate multi-scale emotion features for speech emotion recognition (SER), called Speech Swin-Transformer. Specifically, we first divide the speech spectrogram into segment-level patches in the time domain, composed of multiple frame patches. These segment-level patches are then encoded using a stack of Swin blocks, in which a local window Transformer is utilized to explore local inter-frame emotional information across frame patches of each segment patch. After that, we also design a shifted window Transformer to compensate for patch correlations near the boundaries of segment patches. Finally, we employ a patch merging operation to aggregate segment-level emotional features for hierarchical speech representation by expanding the receptive field of Transformer from frame-level to segment-level. Experimental results demonstrate that our proposed Speech Swin-Transformer outperforms the state-of-the-art methods.