Leveraging Black-box Models to Assess Feature Importance in Unconditional Distribution

Understanding how changes in explanatory features affect the unconditional distribution of the outcome is important in many applications. However, existing black-box predictive models are not readily suited for analyzing such questions. In this work, we develop an approximation method to compute the feature importance curves relevant to the unconditional distribution of outcomes, while leveraging the power of pre-trained black-box predictive models. The feature importance curves measure the changes across quantiles of outcome distribution given an external impact of change in the explanatory features. Through extensive numerical experiments and real data examples, we demonstrate that our approximation method produces sparse and faithful results, and is computationally efficient.

Guided and Fused: Efficient Frozen CLIP-ViT with Feature Guidance and Multi-Stage Feature Fusion for Generalizable Deepfake Detection

The rise of generative models has sparked concerns about image authenticity online, highlighting the urgent need for an effective and general detector. Recent methods leveraging the frozen pre-trained CLIP-ViT model have made great progress in deepfake detection. However, these models often rely on visual-general features directly extracted by the frozen network, which contain excessive information irrelevant to the task, resulting in limited detection performance. To address this limitation, in this paper, we propose an efficient Guided and Fused Frozen CLIP-ViT (GFF), which integrates two simple yet effective modules. The Deepfake-Specific Feature Guidance Module (DFGM) guides the frozen pre-trained model in extracting features specifically for deepfake detection, reducing irrelevant information while preserving its generalization capabilities. The Multi-Stage Fusion Module (FuseFormer) captures low-level and high-level information by fusing features extracted from each stage of the ViT. This dual-module approach significantly improves deepfake detection by fully leveraging CLIP-ViT's inherent advantages. Extensive experiments demonstrate the effectiveness and generalization ability of GFF, which achieves state-of-the-art performance with optimal results in only 5 training epochs. Even when trained on only 4 classes of ProGAN, GFF achieves nearly 99% accuracy on unseen GANs and maintains an impressive 97% accuracy on unseen diffusion models.

Leveraging Web-Crawled Data for High-Quality Fine-Tuning

Most large language models are fine-tuned using either expensive human-annotated data or GPT-4 generated data which cannot guarantee performance in certain domains. We argue that although the web-crawled data often has formatting errors causing semantic inaccuracies, it can still serve as a valuable source for high-quality supervised fine-tuning in specific domains without relying on advanced models like GPT-4. To this end, we create a paired training dataset automatically by aligning web-crawled data with a smaller set of high-quality data. By training a language model on this dataset, we can convert web data with irregular formats into high-quality ones. Our experiments show that training with the model-transformed data yields better results, surpassing training with only high-quality data by an average score of 9.4% in Chinese math problems. Additionally, our 7B model outperforms several open-source models larger than 32B and surpasses well-known closed-source models such as GPT-3.5, highlighting the efficacy of our approach.

Revisiting Modularity Maximization for Graph Clustering: A Contrastive Learning Perspective

Graph clustering, a fundamental and challenging task in graph mining, aims to classify nodes in a graph into several disjoint clusters. In recent years, graph contrastive learning (GCL) has emerged as a dominant line of research in graph clustering and advances the new state-of-the-art. However, GCL-based methods heavily rely on graph augmentations and contrastive schemes, which may potentially introduce challenges such as semantic drift and scalability issues. Another promising line of research involves the adoption of modularity maximization, a popular and effective measure for community detection, as the guiding principle for clustering tasks. Despite the recent progress, the underlying mechanism of modularity maximization is still not well understood. In this work, we dig into the hidden success of modularity maximization for graph clustering. Our analysis reveals the strong connections between modularity maximization and graph contrastive learning, where positive and negative examples are naturally defined by modularity. In light of our results, we propose a community-aware graph clustering framework, coined MAGI, which leverages modularity maximization as a contrastive pretext task to effectively uncover the underlying information of communities in graphs, while avoiding the problem of semantic drift. Extensive experiments on multiple graph datasets verify the effectiveness of MAGI in terms of scalability and clustering performance compared to state-of-the-art graph clustering methods. Notably, MAGI easily scales a sufficiently large graph with 100M nodes while outperforming strong baselines.

A Selective Review on Statistical Methods for Massive Data Computation: Distributed Computing, Subsampling, and Minibatch Techniques

This paper presents a selective review of statistical computation methods for massive data analysis. A huge amount of statistical methods for massive data computation have been rapidly developed in the past decades. In this work, we focus on three categories of statistical computation methods: (1) distributed computing, (2) subsampling methods, and (3) minibatch gradient techniques. The first class of literature is about distributed computing and focuses on the situation, where the dataset size is too huge to be comfortably handled by one single computer. In this case, a distributed computation system with multiple computers has to be utilized. The second class of literature is about subsampling methods and concerns about the situation, where the sample size of dataset is small enough to be placed on one single computer but too large to be easily processed by its memory as a whole. The last class of literature studies those minibatch gradient related optimization techniques, which have been extensively used for optimizing various deep learning models.

Target Speaker Extraction by Directly Exploiting Contextual Information in the Time-Frequency Domain

In target speaker extraction, many studies rely on the speaker embedding which is obtained from an enrollment of the target speaker and employed as the guidance. However, solely using speaker embedding may not fully utilize the contextual information contained in the enrollment. In this paper, we directly exploit this contextual information in the time-frequency (T-F) domain. Specifically, the T-F representations of the enrollment and the mixed signal are interacted to compute the weighting matrices through an attention mechanism. These weighting matrices reflect the similarity among different frames of the T-F representations and are further employed to obtain the consistent T-F representations of the enrollment. These consistent representations are served as the guidance, allowing for better exploitation of the contextual information. Furthermore, the proposed method achieves the state-of-the-art performance on the benchmark dataset and shows its effectiveness in the complex scenarios.

Inferring Intentions to Speak Using Accelerometer Data In-the-Wild

Humans have good natural intuition to recognize when another person has something to say. It would be interesting if an AI can also recognize intentions to speak. Especially in scenarios when an AI is guiding a group discussion, this can be a useful skill. This work studies the inference of successful and unsuccessful intentions to speak from accelerometer data. This is chosen because it is privacy-preserving and feasible for in-the-wild settings since it can be placed in a smart badge. Data from a real-life social networking event is used to train a machine-learning model that aims to infer intentions to speak. A subset of unsuccessful intention-to-speak cases in the data is annotated. The model is trained on the successful intentions to speak and evaluated on both the successful and unsuccessful cases. In conclusion, there is useful information in accelerometer data, but not enough to reliably capture intentions to speak. For example, posture shifts are correlated with intentions to speak, but people also often shift posture without having an intention to speak, or have an intention to speak without shifting their posture. More modalities are likely needed to reliably infer intentions to speak.

LasTGL: An Industrial Framework for Large-Scale Temporal Graph Learning

Over the past few years, graph neural networks (GNNs) have become powerful and practical tools for learning on (static) graph-structure data. However, many real-world applications, such as social networks and e-commerce, involve temporal graphs where nodes and edges are dynamically evolving. Temporal graph neural networks (TGNNs) have progressively emerged as an extension of GNNs to address time-evolving graphs and have gradually become a trending research topic in both academics and industry. Advancing research and application in such an emerging field necessitates the development of new tools to compose TGNN models and unify their different schemes for dealing with temporal graphs. In this work, we introduce LasTGL, an industrial framework that integrates unified and extensible implementations of common temporal graph learning algorithms for various advanced tasks. The purpose of LasTGL is to provide the essential building blocks for solving temporal graph learning tasks, focusing on the guiding principles of user-friendliness and quick prototyping on which PyTorch is based. In particular, LasTGL provides comprehensive temporal graph datasets, TGNN models and utilities along with well-documented tutorials, making it suitable for both absolute beginners and expert deep learning practitioners alike.

On RIS-Aided SIMO Gaussian Channels: Towards A Single-RF MIMO Transceiver Architecture

In this paper, for a single-input multiple-output (SIMO) system aided by a passive reconfigurable intelligent surface (RIS), the joint transmission accomplished by the single transmit antenna and the RIS with multiple controllable reflective elements is considered. Relying on a general capacity upper bound derived by a maximum-trace argument, we respectively characterize the capacity of such \rev{a} channel in the low-SNR or the rank-one regimes, in which the optimal configuration of the RIS is proved to be beamforming with carefully-chosen phase shifts. To exploit the potential of modulating extra information on the RIS, based on the QR decomposition, successive interference cancellation, and a strategy named \textit{partially beamforming and partially information-carrying}, we propose a novel transceiver architecture with only a single RF front end at the transmitter, by which the considered channel can be regarded as a concatenation of a vector Gaussian channel and several phase-modulated channels. Especially, we investigate a class of vector Gaussian channels with a hypersphere input support constraint, and not only generalize the existing result to arbitrary-dimensional real spaces but also present its high-order capacity asymptotics, by which both capacities of hypersphere-constrained channels and achievable rates of the proposed transceiver with two different signaling schemes can be well-approximated. Information-theoretic analyses show that the transceiver architecture designed for the SIMO channel has a boosted multiplexing gain, rather than one for the conventionally-used optimized beamforming scheme.Numerical results verify our derived asymptotics and show notable superiority of the proposed transceiver.

Energy-Efficient Multidimensional Constellation Based on Leech Lattice for Visible Light Communications

In this paper, a 24-dimensional geometrically-shaped constellation design based on Leech lattice is presented for indoor visible light communications (VLCs) with a peak-and an average-intensity input constraints. Firstly, by leveraging tools from large deviation theory, we characterize second-order asymptotics of the optimal constellation shaping region under aforementioned intensity constraints, which further refine our previous results in [Chen. et. al, 2020]. Within the optimal geometrical shaping region, we develop an energy-efficient 24-dimensional constellation design, where a significant coding gain brought by the Leech lattice and the nearly-maximum shaping gain are incorporated by using a strategy called coarsely shaping and finely coding. Fast algorithms for constellation mapping and demodulation are presented as well. Numerical results verifies the superiority of our results as compared with existing methods.