arXiv subjects

Wenhao Wang

Publications and source records attributed to Wenhao Wang.

At least 91 records · Page 5Linked to original sources

arXiv2025

Bellman Error Centering

This paper revisits the recently proposed reward centering algorithms including simple reward centering (SRC) and value-based reward centering (VRC), and points out that SRC is indeed the reward centering, while VRC is essentially Bellman error centering (BEC). Based on BEC, we provide the centered fixpoint for tabular value functions, as well as the centered TD fixpoint for linear value function approximation. We design the on-policy CTD algorithm and the off-policy CTDC algorithm, and prove the convergence of both algorithms. Finally, we experimentally validate the stability of our proposed algorithms. Bellman error centering facilitates the extension to various reinforcement learning algorithms.

cs.LG↗

arXiv2024

Localizing Memorization in SSL Vision Encoders

Recent work on studying memorization in self-supervised learning (SSL) suggests that even though SSL encoders are trained on millions of images, they still memorize individual data points. While effort has been put into characterizing the memorized data and linking encoder memorization to downstream utility, little is known about where the memorization happens inside SSL encoders. To close this gap, we propose two metrics for localizing memorization in SSL encoders on a per-layer (layermem) and per-unit basis (unitmem). Our localization methods are independent of the downstream task, do not require any label information, and can be performed in a forward pass. By localizing memorization in various encoder architectures (convolutional and transformer-based) trained on diverse datasets with contrastive and non-contrastive SSL frameworks, we find that (1) while SSL memorization increases with layer depth, highly memorizing units are distributed across the entire encoder, (2) a significant fraction of units in SSL encoders experiences surprisingly high memorization of individual data points, which is in contrast to models trained under supervision, (3) atypical (or outlier) data points cause much higher layer and unit memorization than standard data points, and (4) in vision transformers, most memorization happens in the fully-connected layers. Finally, we show that localizing memorization in SSL has the potential to improve fine-tuning and to inform pruning strategies.

cs.LG↗

arXiv2024

A Variance Minimization Approach to Temporal-Difference Learning

Fast-converging algorithms are a contemporary requirement in reinforcement learning. In the context of linear function approximation, the magnitude of the smallest eigenvalue of the key matrix is a major factor reflecting the convergence speed. Traditional value-based RL algorithms focus on minimizing errors. This paper introduces a variance minimization (VM) approach for value-based RL instead of error minimization. Based on this approach, we proposed two objectives, the Variance of Bellman Error (VBE) and the Variance of Projected Bellman Error (VPBE), and derived the VMTD, VMTDC, and VMETD algorithms. We provided proofs of their convergence and optimal policy invariance of the variance minimization. Experimental studies validate the effectiveness of the proposed algorithms.

cs.LG↗

arXiv2024

Energy Efficiency Analysis of IRS-aided Wireless Communication Systems Under Statistical QoS Constraints: An Information-Theoretic Perspective

This paper investigates the information-theoretic energy efficiency of intelligent reflecting surface (IRS)-aided wireless communication systems, taking into account the statistical quality-of-service (QoS) constraints on delay violation probabilities. Specifically, effective capacity is adopted to capture the maximum constant arrival rate that can be supported by a time-varying service process while fulfilling these statistical QoS requirements. We derive the minimum bit energy required for the IRS-aided wireless communication system under QoS constraints and analyze the spectral efficiency and energy efficiency tradeoff at low but nonzero signal-to-noise ratio (SNR) levels by also characterizing the wideband slope values. Our analysis demonstrates that the energy efficiency for the considered system under statistical QoS constraints can approach that for a system without QoS limitations in the low-SNR regime. Additionally, deploying a sufficiently large number of practical IRS reflecting elements can substantially reduce energy consumption required to achieve desired spectral efficiency performance in the low-power regime, even with limited bit-resolution phase shifters. Besides, we reveal that compared with the results applied to the low-power regime, higher effective capacity performance can be achieved in scenarios with sparse multipath fading while achieving the same minimum bit energy in the wideband regime.

cs.IT↗

arXiv2024

The Road to Trust: Building Enclaves within Confidential VMs

Integrity is critical for maintaining system security, as it ensures that only genuine software is loaded onto a machine. Although confidential virtual machines (CVMs) function within isolated environments separate from the host, it is important to recognize that users still encounter challenges in maintaining control over the integrity of the code running within the trusted execution environments (TEEs). The presence of a sophisticated operating system (OS) raises the possibility of dynamically creating and executing any code, making user applications within TEEs vulnerable to interference or tampering if the guest OS is compromised. To address this issue, this paper introduces NestedSGX, a framework which leverages virtual machine privilege level (VMPL), a recent hardware feature available on AMD SEV-SNP to enable the creation of hardware enclaves within the guest VM. Similar to Intel SGX, NestedSGX considers the guest OS untrusted for loading potentially malicious code. It ensures that only trusted and measured code executed within the enclave can be remotely attested. To seamlessly protect existing applications, NestedSGX aims for compatibility with Intel SGX by simulating SGX leaf functions. We have also ported the SGX SDK and the Occlum library OS to NestedSGX, enabling the use of existing SGX toolchains and applications in the system. Performance evaluations show that context switches in NestedSGX take about 32,000 -- 34,000 cycles, approximately $1.9\times$ -- $2.1\times$ higher than that of Intel SGX. NestedSGX incurs minimal overhead in most real-world applications, with an average overhead below 2% for computation and memory intensive workloads and below 15.68% for I/O intensive workloads.

cs.CR↗

arXiv2024

KnowledgeSG: Privacy-Preserving Synthetic Text Generation with Knowledge Distillation from Server

The success of large language models (LLMs) facilitate many parties to fine-tune LLMs on their own private data. However, this practice raises privacy concerns due to the memorization of LLMs. Existing solutions, such as utilizing synthetic data for substitution, struggle to simultaneously improve performance and preserve privacy. They either rely on a local model for generation, resulting in a performance decline, or take advantage of APIs, directly exposing the data to API servers. To address this issue, we propose KnowledgeSG, a novel client-server framework which enhances synthetic data quality and improves model performance while ensuring privacy. We achieve this by learning local knowledge from the private data with differential privacy (DP) and distilling professional knowledge from the server. Additionally, inspired by federated learning, we transmit models rather than data between the client and server to prevent privacy leakage. Extensive experiments in medical and financial domains demonstrate the effectiveness of KnowledgeSG. Our code is now publicly available at https://github.com/wwh0411/KnowledgeSG.

cs.CR↗

arXiv2024

Image Copy Detection for Diffusion Models

Images produced by diffusion models are increasingly popular in digital artwork and visual marketing. However, such generated images might replicate content from existing ones and pose the challenge of content originality. Existing Image Copy Detection (ICD) models, though accurate in detecting hand-crafted replicas, overlook the challenge from diffusion models. This motivates us to introduce ICDiff, the first ICD specialized for diffusion models. To this end, we construct a Diffusion-Replication (D-Rep) dataset and correspondingly propose a novel deep embedding method. D-Rep uses a state-of-the-art diffusion model (Stable Diffusion V1.5) to generate 40, 000 image-replica pairs, which are manually annotated into 6 replication levels ranging from 0 (no replication) to 5 (total replication). Our method, PDF-Embedding, transforms the replication level of each image-replica pair into a probability density function (PDF) as the supervision signal. The intuition is that the probability of neighboring replication levels should be continuous and smooth. Experimental results show that PDF-Embedding surpasses protocol-driven methods and non-PDF choices on the D-Rep test set. Moreover, by utilizing PDF-Embedding, we find that the replication ratios of well-known diffusion models against an open-source gallery range from 10% to 20%. The project is publicly available at https://icdiff.github.io/.

cs.CV↗

arXiv2024

VidProM: A Million-scale Real Prompt-Gallery Dataset for Text-to-Video Diffusion Models

The arrival of Sora marks a new era for text-to-video diffusion models, bringing significant advancements in video generation and potential applications. However, Sora, along with other text-to-video diffusion models, is highly reliant on prompts, and there is no publicly available dataset that features a study of text-to-video prompts. In this paper, we introduce VidProM, the first large-scale dataset comprising 1.67 Million unique text-to-Video Prompts from real users. Additionally, this dataset includes 6.69 million videos generated by four state-of-the-art diffusion models, alongside some related data. We initially discuss the curation of this large-scale dataset, a process that is both time-consuming and costly. Subsequently, we underscore the need for a new prompt dataset specifically designed for text-to-video generation by illustrating how VidProM differs from DiffusionDB, a large-scale prompt-gallery dataset for image generation. Our extensive and diverse dataset also opens up many exciting new research areas. For instance, we suggest exploring text-to-video prompt engineering, efficient video generation, and video copy detection for diffusion models to develop better, more efficient, and safer models. The project (including the collected dataset VidProM and related code) is publicly available at https://vidprom.github.io under the CC-BY-NC 4.0 License.

cs.CV↗

arXiv2024

AnyPattern: Towards In-context Image Copy Detection

This paper explores in-context learning for image copy detection (ICD), i.e., prompting an ICD model to identify replicated images with new tampering patterns without the need for additional training. The prompts (or the contexts) are from a small set of image-replica pairs that reflect the new patterns and are used at inference time. Such in-context ICD has good realistic value, because it requires no fine-tuning and thus facilitates fast reaction against the emergence of unseen patterns. To accommodate the "seen $\rightarrow$ unseen" generalization scenario, we construct the first large-scale pattern dataset named AnyPattern, which has the largest number of tamper patterns ($90$ for training and $10$ for testing) among all the existing ones. We benchmark AnyPattern with popular ICD methods and reveal that existing methods barely generalize to novel patterns. We further propose a simple in-context ICD method named ImageStacker. ImageStacker learns to select the most representative image-replica pairs and employs them as the pattern prompts in a stacking manner (rather than the popular concatenation manner). Experimental results show (1) training with our large-scale dataset substantially benefits pattern generalization ($+26.66 \%$ $μAP$), (2) the proposed ImageStacker facilitates effective in-context ICD (another round of $+16.75 \%$ $μAP$), and (3) AnyPattern enables in-context ICD, i.e., without such a large-scale dataset, in-context learning does not emerge even with our ImageStacker. Beyond the ICD task, we also demonstrate how AnyPattern can benefit artists, i.e., the pattern retrieval method trained on AnyPattern can be generalized to identify style mimicry by text-to-image models. The project is publicly available at https://anypattern.github.io.

cs.CV↗

arXiv2024

MonoFormer: One Transformer for Both Diffusion and Autoregression

Most existing multimodality methods use separate backbones for autoregression-based discrete text generation and diffusion-based continuous visual generation, or the same backbone by discretizing the visual data to use autoregression for both text and visual generation. In this paper, we propose to study a simple idea: share one transformer for both autoregression and diffusion. The feasibility comes from two main aspects: (i) Transformer is successfully applied to diffusion for visual generation, and (ii) transformer training for autoregression and diffusion is very similar, and the difference merely lies in that diffusion uses bidirectional attention mask and autoregression uses causal attention mask. Experimental results show that our approach achieves comparable image generation performance to current state-of-the-art methods as well as maintains the text generation capability. The project is publicly available at https://monoformer.github.io/.

cs.CV↗

arXiv2024

FedRSU: Federated Learning for Scene Flow Estimation on Roadside Units

Roadside unit (RSU) can significantly improve the safety and robustness of autonomous vehicles through Vehicle-to-Everything (V2X) communication. Currently, the usage of a single RSU mainly focuses on real-time inference and V2X collaboration, while neglecting the potential value of the high-quality data collected by RSU sensors. Integrating the vast amounts of data from numerous RSUs can provide a rich source of data for model training. However, the absence of ground truth annotations and the difficulty of transmitting enormous volumes of data are two inevitable barriers to fully exploiting this hidden value. In this paper, we introduce FedRSU, an innovative federated learning framework for self-supervised scene flow estimation. In FedRSU, we present a recurrent self-supervision training paradigm, where for each RSU, the scene flow prediction of points at every timestamp can be supervised by its subsequent future multi-modality observation. Another key component of FedRSU is federated learning, where multiple devices collaboratively train an ML model while keeping the training data local and private. With the power of the recurrent self-supervised learning paradigm, FL is able to leverage innumerable underutilized data from RSU. To verify the FedRSU framework, we construct a large-scale multi-modality dataset RSU-SF. The dataset consists of 17 RSU clients, covering various scenarios, modalities, and sensor settings. Based on RSU-SF, we show that FedRSU can greatly improve model performance in ITS and provide a comprehensive benchmark under diverse FL scenarios. To the best of our knowledge, we provide the first real-world LiDAR-camera multi-modal dataset and benchmark for the FL community.

cs.CV↗

arXiv2024

From Ideal Membership Problem for polynomial rings to Dehn Functions of Metabelian Groups

The ideal membership problem asks whether an element in the ring belongs to the given ideal. In this paper, we propose a function that reflecting the complexity of the ideal membership problem in the ring of Laurent polynomials with integer coefficients. We also connect the complexity function we define to the Dehn function of a metabelian group, in the hope of constructing a metabelian group with superexponential Dehn function.

math.GR↗

arXiv2024

Computational Investigation on the formation of liquid-fueled oblique detonation waves

Utilizing a two-phase supersonic chemically reacting flow solver with the Eulerian-Lagrangian method implemented in OpenFOAM, this study computationally investigates the formation of liquid-fueled oblique detonation waves (ODWs) within a pre-injection oblique detonation wave engine operating at an altitude of 30 km and a velocity of Mach 9. The inflow undergoes two-stage compression, followed by uniform mixing with randomly distributed n-heptane droplets before entering the combustor. The study examines the effects of droplet breakup models, gas-liquid ratios, and on-wedge strips on the ODW formation. Results indicate that under the pure-droplet condition, the ODW fails to form within the combustor, irrespective of the breakup models used. However, increasing the proportion of n-heptane vapor in the fuel/air mixture facilitates the ODW formation, because the n-heptane vapor rapidly participates in the gaseous reactions, producing heat and accelerating the transition from low- to intermediate-temperature chemistry. Additionally, the presence of on-wedge strips enhances ODW formation by inducing a bow shock wave within the combustor, which significantly increases the temperature, directly triggering intermediate-temperature chemistry and subsequent heat-release reactions, thereby facilitating the formation of ODW.

physics.flu-dyn↗

arXiv2024

Cabin: Confining Untrusted Programs within Confidential VMs

Confidential computing safeguards sensitive computations from untrusted clouds, with Confidential Virtual Machines (CVMs) providing a secure environment for guest OS. However, CVMs often come with large and vulnerable operating system kernels, making them susceptible to attacks exploiting kernel weaknesses. The imprecise control over the read/write access in the page table has allowed attackers to exploit vulnerabilities. The lack of security hierarchy leads to insufficient separation between untrusted applications and guest OS, making the kernel susceptible to direct threats from untrusted programs. This study proposes Cabin, an isolated execution framework within guest VM utilizing the latest AMD SEV-SNP technology. Cabin shields untrusted processes to the user space of a lower virtual machine privilege level (VMPL) by introducing a proxy-kernel between the confined processes and the guest OS. Furthermore, we propose execution protection mechanisms based on fine-gained control of VMPL privilege for vulnerable programs and the proxy-kernel to minimize the attack surface. We introduce asynchronous forwarding mechanism and anonymous memory management to reduce the performance impact. The evaluation results show that the Cabin framework incurs a modest overhead (5% on average) on Nbench and WolfSSL benchmarks.

cs.CR↗

arXiv2024

Memorization in Self-Supervised Learning Improves Downstream Generalization

Self-supervised learning (SSL) has recently received significant attention due to its ability to train high-performance encoders purely on unlabeled data-often scraped from the internet. This data can still be sensitive and empirical evidence suggests that SSL encoders memorize private information of their training data and can disclose them at inference time. Since existing theoretical definitions of memorization from supervised learning rely on labels, they do not transfer to SSL. To address this gap, we propose SSLMem, a framework for defining memorization within SSL. Our definition compares the difference in alignment of representations for data points and their augmented views returned by both encoders that were trained on these data points and encoders that were not. Through comprehensive empirical analysis on diverse encoder architectures and datasets we highlight that even though SSL relies on large datasets and strong augmentations-both known in supervised learning as regularization techniques that reduce overfitting-still significant fractions of training data points experience high memorization. Through our empirical results, we show that this memorization is essential for encoders to achieve higher generalization performance on different downstream tasks.

cs.LG↗

arXiv2024

A Dynamic Droplet Breakup Model for Eulerian-Lagrangian Simulation of Liquid-fueled Detonation

This study proposes a dynamic model to reflect the physical image of the droplet breakup process in two-phase detonation flows. This breakup model is implemented in a two-phase detonation solver developed based on an open-source computational fluid dynamic platform, OpenFOAM, and compared with three prevalent models (TAB, PilchErdman, and ReitzKH-RT model) under different droplet diameters in one- and two-dimensional detonation problems. The simulating results show that the present breakup model well predicts experimentally determined detonation parameters such as detonation velocities and post-wave temperature. In addition, the present model has the advantage of being free of the KH breakup time parameter, which is needed by the ReitzKH-RT model to fit the experimental data. The one-dimensional detonation simulations indicate that different breakup models have a slight impact on the detonation wave velocity because the droplet breakup process does not significantly affect the total heat release as long as it is sufficiently fast to sustain the detonation. However, the two-dimensional detonation simulations show that both the breakup model and the droplet initial diameter significantly affect the detonation cell size due to the different droplet distributions predicted by different models. The breakup length, which is the distance from the shock wave to the location at which sufficiently small child droplets appear, affects the chemical reaction zone thickness, which in turn affects the detonation cell size. A longer breakup length will result in a larger detonation cell size.

physics.flu-dyn↗

arXiv2024

TRAD: Enhancing LLM Agents with Step-Wise Thought Retrieval and Aligned Decision

Numerous large language model (LLM) agents have been built for different tasks like web navigation and online shopping due to LLM's wide knowledge and text-understanding ability. Among these works, many of them utilize in-context examples to achieve generalization without the need for fine-tuning, while few of them have considered the problem of how to select and effectively utilize these examples. Recently, methods based on trajectory-level retrieval with task meta-data and using trajectories as in-context examples have been proposed to improve the agent's overall performance in some sequential decision making tasks. However, these methods can be problematic due to plausible examples retrieved without task-specific state transition dynamics and long input with plenty of irrelevant context. In this paper, we propose a novel framework (TRAD) to address these issues. TRAD first conducts Thought Retrieval, achieving step-level demonstration selection via thought matching, leading to more helpful demonstrations and less irrelevant input noise. Then, TRAD introduces Aligned Decision, complementing retrieved demonstration steps with their previous or subsequent steps, which enables tolerance for imperfect thought and provides a choice for balance between more context and less noise. Extensive experiments on ALFWorld and Mind2Web benchmarks show that TRAD not only outperforms state-of-the-art models but also effectively helps in reducing noise and promoting generalization. Furthermore, TRAD has been deployed in real-world scenarios of a global business insurance company and improves the success rate of robotic process automation.

cs.AI↗

arXiv2024

OpenFedLLM: Training Large Language Models on Decentralized Private Data via Federated Learning

Trained on massive publicly available data, large language models (LLMs) have demonstrated tremendous success across various fields. While more data contributes to better performance, a disconcerting reality is that high-quality public data will be exhausted in a few years. In this paper, we offer a potential next step for contemporary LLMs: collaborative and privacy-preserving LLM training on the underutilized distributed private data via federated learning (FL), where multiple data owners collaboratively train a shared model without transmitting raw data. To achieve this, we build a concise, integrated, and research-friendly framework/codebase, named OpenFedLLM. It covers federated instruction tuning for enhancing instruction-following capability, federated value alignment for aligning with human values, and 7 representative FL algorithms. Besides, OpenFedLLM supports training on diverse domains, where we cover 8 training datasets; and provides comprehensive evaluations, where we cover 30+ evaluation metrics. Through extensive experiments, we observe that all FL algorithms outperform local training on training LLMs, demonstrating a clear performance improvement across a variety of settings. Notably, in a financial benchmark, Llama2-7B fine-tuned by applying any FL algorithm can outperform GPT-4 by a significant margin while the model obtained through individual training cannot, demonstrating strong motivation for clients to participate in FL. The code is available at https://github.com/rui-ye/OpenFedLLM.

cs.LG↗