Mixture-of-Experts (MoE) architectures in large language models (LLMs) achieve exceptional performance, but face prohibitive storage and memory requirements. To address these challenges, we present $D^2$-MoE, a new delta decompression compressor for reducing the parameters of MoE LLMs. Based on observations of expert diversity, we decompose their weights into a shared base weight and unique delta weights. Specifically, our method first merges each expert's weight into the base weight using the Fisher information matrix to capture shared components. Then, we compress delta weights through Singular Value Decomposition (SVD) by exploiting their low-rank properties. Finally, we introduce a semi-dynamical structured pruning strategy for the base weights, combining static and dynamic redundancy analysis to achieve further parameter reduction while maintaining input adaptivity. In this way, our $D^2$-MoE successfully compact MoE LLMs to high compression ratios without additional training. Extensive experiments highlight the superiority of our approach, with over 13% performance gains than other compressors on Mixtral|Phi-3.5|DeepSeek|Qwen2 MoE LLMs at 40$\sim$60% compression rates. Codes are available in https://github.com/lliai/D2MoE.

Rapid advancements in speech synthesis and voice conversion bring convenience but also new security risks, creating an urgent need for effective audio deepfake detection. Although current models perform well, their effectiveness diminishes when confronted with the diverse and evolving nature of real-world deepfakes. To address this issue, we propose a continual learning method named Region-Based Optimization (RegO) for audio deepfake detection. Specifically, we use the Fisher information matrix to measure important neuron regions for real and fake audio detection, dividing them into four regions. First, we directly fine-tune the less important regions to quickly adapt to new tasks. Next, we apply gradient optimization in parallel for regions important only to real audio detection, and in orthogonal directions for regions important only to fake audio detection. For regions that are important to both, we use sample proportion-based adaptive gradient optimization. This region-adaptive optimization ensures an appropriate trade-off between memory stability and learning plasticity. Additionally, to address the increase of redundant neurons from old tasks, we further introduce the Ebbinghaus forgetting mechanism to release them, thereby promoting the capability of the model to learn more generalized discriminative features. Experimental results show our method achieves a 21.3% improvement in EER over the state-of-the-art continual learning approach RWM for audio deepfake detection. Moreover, the effectiveness of RegO extends beyond the audio deepfake detection domain, showing potential significance in other tasks, such as image recognition. The code is available at https://github.com/cyjie429/RegO

Real-world recommendation systems commonly offer diverse content scenarios for users to interact with. Considering the enormous number of users in industrial platforms, it is infeasible to utilize a single unified recommendation model to meet the requirements of all scenarios. Usually, separate recommendation pipelines are established for each distinct scenario. This practice leads to challenges in comprehensively grasping users' interests. Recent research endeavors have been made to tackle this problem by pre-training models to encapsulate the overall interests of users. Traditional pre-trained recommendation models mainly capture user interests by leveraging collaborative signals. Nevertheless, a prevalent drawback of these systems is their incapacity to handle long-tail items and cold-start scenarios. With the recent advent of large language models, there has been a significant increase in research efforts focused on exploiting LLMs to extract semantic information for users and items. However, text-based recommendations highly rely on elaborate feature engineering and frequently fail to capture collaborative similarities. To overcome these limitations, we propose a novel pre-training framework for sequential recommendation, termed PRECISE. This framework combines collaborative signals with semantic information. Moreover, PRECISE employs a learning framework that initially models users' comprehensive interests across all recommendation scenarios and subsequently concentrates on the specific interests of target-scene behaviors. We demonstrate that PRECISE precisely captures the entire range of user interests and effectively transfers them to the target interests. Empirical findings reveal that the PRECISE framework attains outstanding performance on both public and industrial datasets.

Open environment oriented open set model attribution of deepfake audio is an emerging research topic, aiming to identify the generation models of deepfake audio. Most previous work requires manually setting a rejection threshold for unknown classes to compare with predicted probabilities. However, models often overfit training instances and generate overly confident predictions. Moreover, thresholds that effectively distinguish unknown categories in the current dataset may not be suitable for identifying known and unknown categories in another data distribution. To address the issues, we propose a novel framework for open set model attribution of deepfake audio with rejection threshold adaptation (ReTA). Specifically, the reconstruction error learning module trains by combining the representation of system fingerprints with labels corresponding to either the target class or a randomly chosen other class label. This process generates matching and non-matching reconstructed samples, establishing the reconstruction error distributions for each class and laying the foundation for the reject threshold calculation module. The reject threshold calculation module utilizes gaussian probability estimation to fit the distributions of matching and non-matching reconstruction errors. It then computes adaptive reject thresholds for all classes through probability minimization criteria. The experimental results demonstrate the effectiveness of ReTA in improving the open set model attributes of deepfake audio.

Multimodal emotion recognition is an important research topic in artificial intelligence. Over the past few decades, researchers have made remarkable progress by increasing dataset size and building more effective architectures. However, due to various reasons (such as complex environments and inaccurate annotations), current systems are hard to meet the demands of practical applications. Therefore, we organize a series of challenges around emotion recognition to further promote the development of this area. Last year, we launched MER2023, focusing on three topics: multi-label learning, noise robustness, and semi-supervised learning. This year, we continue to organize MER2024. In addition to expanding the dataset size, we introduce a new track around open-vocabulary emotion recognition. The main consideration for this track is that existing datasets often fix the label space and use majority voting to enhance annotator consistency, but this process may limit the model's ability to describe subtle emotions. In this track, we encourage participants to generate any number of labels in any category, aiming to describe the emotional state as accurately as possible. Our baseline is based on MERTools and the code is available at: https://github.com/zeroQiaoba/MERTools/tree/master/MER2024.

The rapid progress of deep speech synthesis models has posed significant threats to society such as malicious content manipulation. Therefore, many studies have emerged to detect the so-called deepfake audio. However, existing works focus on the binary detection of real audio and fake audio. In real-world scenarios such as model copyright protection and digital evidence forensics, it is needed to know what tool or model generated the deepfake audio to explain the decision. This motivates us to ask: Can we recognize the system fingerprints of deepfake audio? In this paper, we present the first deepfake audio dataset for system fingerprint recognition (SFR) and conduct an initial investigation. We collected the dataset from the speech synthesis systems of seven Chinese vendors that use the latest state-of-the-art deep learning technologies, including both clean and compressed sets. In addition, to facilitate the further development of system fingerprint recognition methods, we provide extensive benchmarks that can be compared and research findings. The dataset will be publicly available. .