MusER: Musical Element-Based Regularization for Generating Symbolic Music with Emotion

Generating music with emotion is an important task in automatic music generation, in which emotion is evoked through a variety of musical elements (such as pitch and duration) that change over time and collaborate with each other. However, prior research on deep learning-based emotional music generation has rarely explored the contribution of different musical elements to emotions, let alone the deliberate manipulation of these elements to alter the emotion of music, which is not conducive to fine-grained element-level control over emotions. To address this gap, we present a novel approach employing musical element-based regularization in the latent space to disentangle distinct elements, investigate their roles in distinguishing emotions, and further manipulate elements to alter musical emotions. Specifically, we propose a novel VQ-VAE-based model named MusER. MusER incorporates a regularization loss to enforce the correspondence between the musical element sequences and the specific dimensions of latent variable sequences, providing a new solution for disentangling discrete sequences. Taking advantage of the disentangled latent vectors, a two-level decoding strategy that includes multiple decoders attending to latent vectors with different semantics is devised to better predict the elements. By visualizing latent space, we conclude that MusER yields a disentangled and interpretable latent space and gain insights into the contribution of distinct elements to the emotional dimensions (i.e., arousal and valence). Experimental results demonstrate that MusER outperforms the state-of-the-art models for generating emotional music in both objective and subjective evaluation. Besides, we rearrange music through element transfer and attempt to alter the emotion of music by transferring emotion-distinguishable elements.

High-Fidelity Face Swapping with Style Blending

Face swapping has gained significant traction, driven by the plethora of human face synthesis facilitated by deep learning methods. However, previous face swapping methods that used generative adversarial networks (GANs) as backbones have faced challenges such as inconsistency in blending, distortions, artifacts, and issues with training stability. To address these limitations, we propose an innovative end-to-end framework for high-fidelity face swapping. First, we introduce a StyleGAN-based facial attributes encoder that extracts essential features from faces and inverts them into a latent style code, encapsulating indispensable facial attributes for successful face swapping. Second, we introduce an attention-based style blending module to effectively transfer Face IDs from source to target. To ensure accurate and quality transferring, a series of constraint measures including contrastive face ID learning, facial landmark alignment, and dual swap consistency is implemented. Finally, the blended style code is translated back to the image space via the style decoder, which is of high training stability and generative capability. Extensive experiments on the CelebA-HQ dataset highlight the superior visual quality of generated images from our face-swapping methodology when compared to other state-of-the-art methods, and the effectiveness of each proposed module. Source code and weights will be publicly available.

Self-Supervised Disentangled Representation Learning for Robust Target Speech Extraction

Speech signals are inherently complex as they encompass both global acoustic characteristics and local semantic information. However, in the task of target speech extraction, certain elements of global and local semantic information in the reference speech, which are irrelevant to speaker identity, can lead to speaker confusion within the speech extraction network. To overcome this challenge, we propose a self-supervised disentangled representation learning method. Our approach tackles this issue through a two-phase process, utilizing a reference speech encoding network and a global information disentanglement network to gradually disentangle the speaker identity information from other irrelevant factors. We exclusively employ the disentangled speaker identity information to guide the speech extraction network. Moreover, we introduce the adaptive modulation Transformer to ensure that the acoustic representation of the mixed signal remains undisturbed by the speaker embeddings. This component incorporates speaker embeddings as conditional information, facilitating natural and efficient guidance for the speech extraction network. Experimental results substantiate the effectiveness of our meticulously crafted approach, showcasing a substantial reduction in the likelihood of speaker confusion.

CASR: Refining Action Segmentation via Magrinalizing Frame-levle Causal Relationships

Integrating deep learning and causal discovery has increased the interpretability of Temporal Action Segmentation (TAS) tasks. However, frame-level causal relationships exist many complicated noises outside the segment-level, making it infeasible to directly express macro action semantics. Thus, we propose Causal Abstraction Segmentation Refiner (CASR), which can refine TAS results from various models by enhancing video causality in marginalizing frame-level casual relationships. Specifically, we define the equivalent frame-level casual model and segment-level causal model, so that the causal adjacency matrix constructed from marginalized frame-level causal relationships has the ability to represent the segmnet-level causal relationships. CASR works out by reducing the difference in the causal adjacency matrix between we constructed and pre-segmentation results of backbone models. In addition, we propose a novel evaluation metric Causal Edit Distance (CED) to evaluate the causal interpretability. Extensive experimental results on mainstream datasets indicate that CASR significantly surpasses existing various methods in action segmentation performance, as well as in causal explainability and generalization.

Holistic Analysis of Hallucination in GPT-4V: Bias and Interference Challenges

While GPT-4V(ision) impressively models both visual and textual information simultaneously, it's hallucination behavior has not been systematically assessed. To bridge this gap, we introduce a new benchmark, namely, the Bias and Interference Challenges in Visual Language Models (Bingo). This benchmark is designed to evaluate and shed light on the two common types of hallucinations in visual language models: bias and interference. Here, bias refers to the model's tendency to hallucinate certain types of responses, possibly due to imbalance in its training data. Interference pertains to scenarios where the judgment of GPT-4V(ision) can be disrupted due to how the text prompt is phrased or how the input image is presented. We identify a notable regional bias, whereby GPT-4V(ision) is better at interpreting Western images or images with English writing compared to images from other countries or containing text in other languages. Moreover, GPT-4V(ision) is vulnerable to leading questions and is often confused when interpreting multiple images together. Popular mitigation approaches, such as self-correction and chain-of-thought reasoning, are not effective in resolving these challenges. We also identified similar biases and interference vulnerabilities with LLaVA and Bard. Our results characterize the hallucination challenges in GPT-4V(ision) and state-of-the-art visual-language models, and highlight the need for new solutions. The Bingo benchmark is available at https://github.com/gzcch/Bingo.

A Review and Roadmap of Deep Causal Model from Different Causal Structures and Representations

The fusion of causal models with deep learning introducing increasingly intricate data sets, such as the causal associations within images or between textual components, has surfaced as a focal research area. Nonetheless, the broadening of original causal concepts and theories to such complex, non-statistical data has been met with serious challenges. In response, our study proposes redefinitions of causal data into three distinct categories from the standpoint of causal structure and representation: definite data, semi-definite data, and indefinite data. Definite data chiefly pertains to statistical data used in conventional causal scenarios, while semi-definite data refers to a spectrum of data formats germane to deep learning, including time-series, images, text, and others. Indefinite data is an emergent research sphere inferred from the progression of data forms by us. To comprehensively present these three data paradigms, we elaborate on their formal definitions, differences manifested in datasets, resolution pathways, and development of research. We summarize key tasks and achievements pertaining to definite and semi-definite data from myriad research undertakings, present a roadmap for indefinite data, beginning with its current research conundrums. Lastly, we classify and scrutinize the key datasets presently utilized within these three paradigms.

SSL Framework for Causal Inconsistency between Structures and Representations

The cross-pollination of deep learning and causal discovery has catalyzed a burgeoning field of research seeking to elucidate causal relationships within non-statistical data forms like images, videos, and text. Such data, often being named `indefinite data', exhibit unique challenges-inconsistency between causal structure and representation, which are not common in conventional data forms. To tackle this issue, we theoretically develop intervention strategies suitable for indefinite data and derive causal consistency condition (CCC). Moreover, we design a self-supervised learning (SSL) framework that considers interventions as `views' and CCC as a `philosophy' with two implement examples on Supervised Specialized Models (SSMs) and Large Language Models (LLMs), respectively. To evaluate pure inconsistency manifestations, we have prepared the first high-quality causal dialogue dataset-Causalogue. Evaluations are also performed on three other downstream tasks. Extensive experimentation has substantiated the efficacy of our methodology, illuminating how CCC could potentially play an influential role in various fields.

Manipulation Mask Generator: High-Quality Image Manipulation Mask Generation Method Based on Modified Total Variation Noise Reduction

In artificial intelligence, any model that wants to achieve a good result is inseparable from a large number of high-quality data. It is especially true in the field of tamper detection. This paper proposes a modified total variation noise reduction method to acquire high-quality tampered images. We automatically crawl original and tampered images from the Baidu PS Bar. Baidu PS Bar is a website where net friends post countless tampered images. Subtracting the original image with the tampered image can highlight the tampered area. However, there is also substantial noise on the final print, so these images can't be directly used in the deep learning model. Our modified total variation noise reduction method is aimed at solving this problem. Because a lot of text is slender, it is easy to lose text information after the opening and closing operation. We use MSER (Maximally Stable Extremal Regions) and NMS (Non-maximum Suppression) technology to extract text information. And then use the modified total variation noise reduction technology to process the subtracted image. Finally, we can obtain an image with little noise by adding the image and text information. And the idea also largely retains the text information. Datasets generated in this way can be used in deep learning models, and they will help the model achieve better results.

Can large language models provide useful feedback on research papers? A large-scale empirical analysis

Expert feedback lays the foundation of rigorous research. However, the rapid growth of scholarly production and intricate knowledge specialization challenge the conventional scientific feedback mechanisms. High-quality peer reviews are increasingly difficult to obtain. Researchers who are more junior or from under-resourced settings have especially hard times getting timely feedback. With the breakthrough of large language models (LLM) such as GPT-4, there is growing interest in using LLMs to generate scientific feedback on research manuscripts. However, the utility of LLM-generated feedback has not been systematically studied. To address this gap, we created an automated pipeline using GPT-4 to provide comments on the full PDFs of scientific papers. We evaluated the quality of GPT-4's feedback through two large-scale studies. We first quantitatively compared GPT-4's generated feedback with human peer reviewer feedback in 15 Nature family journals (3,096 papers in total) and the ICLR machine learning conference (1,709 papers). The overlap in the points raised by GPT-4 and by human reviewers (average overlap 30.85% for Nature journals, 39.23% for ICLR) is comparable to the overlap between two human reviewers (average overlap 28.58% for Nature journals, 35.25% for ICLR). The overlap between GPT-4 and human reviewers is larger for the weaker papers. We then conducted a prospective user study with 308 researchers from 110 US institutions in the field of AI and computational biology to understand how researchers perceive feedback generated by our GPT-4 system on their own papers. Overall, more than half (57.4%) of the users found GPT-4 generated feedback helpful/very helpful and 82.4% found it more beneficial than feedback from at least some human reviewers. While our findings show that LLM-generated feedback can help researchers, we also identify several limitations.

MemDA: Forecasting Urban Time Series with Memory-based Drift Adaptation

Urban time series data forecasting featuring significant contributions to sustainable development is widely studied as an essential task of the smart city. However, with the dramatic and rapid changes in the world environment, the assumption that data obey Independent Identically Distribution is undermined by the subsequent changes in data distribution, known as concept drift, leading to weak replicability and transferability of the model over unseen data. To address the issue, previous approaches typically retrain the model, forcing it to fit the most recent observed data. However, retraining is problematic in that it leads to model lag, consumption of resources, and model re-invalidation, causing the drift problem to be not well solved in realistic scenarios. In this study, we propose a new urban time series prediction model for the concept drift problem, which encodes the drift by considering the periodicity in the data and makes on-the-fly adjustments to the model based on the drift using a meta-dynamic network. Experiments on real-world datasets show that our design significantly outperforms state-of-the-art methods and can be well generalized to existing prediction backbones by reducing their sensitivity to distribution changes.