arXiv subjects

Zihao Wang

Publications and source records attributed to Zihao Wang.

At least 235 records · Page 13Linked to original sources

arXiv2024

Real-Time Machine-Learning-Based Optimization Using Input Convex Long Short-Term Memory Network

Neural network-based optimization and control methods, often referred to as black-box approaches, are increasingly gaining attention in energy and manufacturing systems, particularly in situations where first-principles models are either unavailable or inaccurate. However, their non-convex nature significantly slows down the optimization and control processes, limiting their application in real-time decision-making processes. To address this challenge, we propose a novel Input Convex Long Short-Term Memory (IC-LSTM) network to enhance the computational efficiency of neural network-based optimization. Through two case studies employing real-time neural network-based optimization for optimizing energy and chemical systems, we demonstrate the superior performance of IC-LSTM-based optimization in terms of runtime. Specifically, in a real-time optimization problem of a real-world solar photovoltaic energy system at LHT Holdings in Singapore, IC-LSTM-based optimization achieved at least 4-fold speedup compared to conventional LSTM-based optimization. These results highlight the potential of IC-LSTM networks to significantly enhance the efficiency of neural network-based optimization and control in practical applications. Source code is available at https://github.com/killingbear999/ICLSTM.

cs.LG↗

arXiv2024

MetaBGM: Dynamic Soundtrack Transformation For Continuous Multi-Scene Experiences With Ambient Awareness And Personalization

This paper introduces MetaBGM, a groundbreaking framework for generating background music that adapts to dynamic scenes and real-time user interactions. We define multi-scene as variations in environmental contexts, such as transitions in game settings or movie scenes. To tackle the challenge of converting backend data into music description texts for audio generation models, MetaBGM employs a novel two-stage generation approach that transforms continuous scene and user state data into these texts, which are then fed into an audio generation model for real-time soundtrack creation. Experimental results demonstrate that MetaBGM effectively generates contextually relevant and dynamic background music for interactive applications.

cs.SD↗

arXiv2024

Nanometric dual-comb ranging using photon-level microcavity solitons

Absolute distance measurement with low return power, fast measurement speed, high precision, and immunity to intensity fluctuations is highly demanded in nanotechnology. However, achieving all these objectives simultaneously remains a significant challenge for miniaturized systems. Here, we demonstrate dual-comb ranging (DCR) that encompasses all these capabilities by using counter-propagating (CP) solitons generated in an integrated Si$_3$N$_4$ microresonator. We derive equations linking the DCR precision with comb line powers, revealing the advantage of microcomb's large line spacing in precise ranging. Leveraging the advantage, our system reaches 1-nm-precision and measures nm-scale vibration at frequencies up to 0.9 MHz. We also show that precise DCR is possible even in the presence of strong intensity noise and loss, using a mean received photon number as low as 5.5$\times$10$^{-4}$ per pulse. Our work establishes an optimization principle for dual-comb systems and bridges high performance ranging with foundry-manufactured photonic chips.

physics.optics↗

arXiv2024

A UAV-Enabled Time-Sensitive Data Collection Scheme for Grassland Monitoring Edge Networks

Grassland monitoring is essential for the sustainable development of grassland resources. Traditional Internet of Things (IoT) devices generate critical ecological data, making data loss unacceptable, but the harsh environment complicates data collection. Unmanned Aerial Vehicle (UAV) and mobile edge computing (MEC) offer efficient data collection solutions, enhancing performance on resource-limited mobile devices. In this context, this paper is the first to investigate a UAV-enabled time-sensitive data collection problem (TSDCMP) within grassland monitoring edge networks (GMENs). Unlike many existing data collection scenarios, this problem has three key challenges. First, the total amount of data collected depends significantly on the data collection duration and arrival time of UAV at each access point (AP). Second, the volume of data at different APs varies among regions due to differences in monitoring objects and vegetation coverage. Third, the service requests time and locations from APs are often not adjacent topologically. To address these issues, We formulate the TSDCMP for UAV-enabled GMENs as a mixed-integer programming model in a single trip. This model considers constraints such as the limited energy of UAV, the coupled routing and time scheduling, and the state of APs and UAV arrival time. Subsequently, we propose a novel cooperative heuristic algorithm based on temporal-spatial correlations (CHTSC) that integrates a modified dynamic programming (MDP) into an iterated local search to solve the TSDCMP for UAV-enabled GMENs. This approach fully takes into account the temporal and spatial relationships between consecutive service requests from APs. Systematic simulation studies demonstrate that the mixed-integer programming model effectively represents the TSDCMP within UAV-enabled GMENs.

cs.NI↗

arXiv2024

PSNE: Efficient Spectral Sparsification Algorithms for Scaling Network Embedding

Network embedding has numerous practical applications and has received extensive attention in graph learning, which aims at mapping vertices into a low-dimensional and continuous dense vector space by preserving the underlying structural properties of the graph. Many network embedding methods have been proposed, among which factorization of the Personalized PageRank (PPR for short) matrix has been empirically and theoretically well supported recently. However, several fundamental issues cannot be addressed. (1) Existing methods invoke a seminal Local Push subroutine to approximate \textit{a single} row or column of the PPR matrix. Thus, they have to execute $n$ ($n$ is the number of nodes) Local Push subroutines to obtain a provable PPR matrix, resulting in prohibitively high computational costs for large $n$. (2) The PPR matrix has limited power in capturing the structural similarity between vertices, leading to performance degradation. To overcome these dilemmas, we propose PSNE, an efficient spectral s\textbf{P}arsification method for \textbf{S}caling \textbf{N}etwork \textbf{E}mbedding, which can fast obtain the embedding vectors that retain strong structural similarities. Specifically, PSNE first designs a matrix polynomial sparser to accelerate the calculation of the PPR matrix, which has a theoretical guarantee in terms of the Frobenius norm. Subsequently, PSNE proposes a simple but effective multiple-perspective strategy to enhance further the representation power of the obtained approximate PPR matrix. Finally, PSNE applies a randomized singular value decomposition algorithm on the sparse and multiple-perspective PPR matrix to get the target embedding vectors. Experimental evaluation of real-world and synthetic datasets shows that our solutions are indeed more efficient, effective, and scalable compared with ten competitors.

cs.LG↗

arXiv2024

A microcomb-empowered Fourier domain mode-locked LiDAR

Light detection and ranging (LiDAR) has emerged as an indispensable tool in autonomous technology. Among its various techniques, frequency modulated continuous wave (FMCW) LiDAR stands out due to its capability to operate with ultralow return power, immunity to unwanted light, and simultaneous acquisition of distance and velocity. However, achieving a rapid update rate with sub-micron precision remains a challenge for FMCW LiDARs. Here, we present such a LiDAR with a sub-10 nm precision and a 24.6 kHz update rate by combining a broadband Fourier domain mode-locked (FDML) laser with a silicon nitride soliton microcomb. An ultrahigh frequency chirp rate up to 320 PHz/s is linearized by a 50 GHz microcomb to reach this performance. Our theoretical analysis also contributes to resolving the challenge of FMCW velocity measurements with nonlinear frequency sweeps and enables us to realize velocity measurement with an uncertainty below 0.4 mm/s. Our work shows how nanophotonic microcombs can unlock the potential of ultrafast frequency sweeping lasers for applications including LiDAR, optical coherence tomography and sensing.

physics.optics↗

arXiv2024

REMAST: Real-time Emotion-based Music Arrangement with Soft Transition

Music as an emotional intervention medium has important applications in scenarios such as music therapy, games, and movies. However, music needs real-time arrangement according to changing emotions, bringing challenges to balance emotion real-time fit and soft emotion transition due to the fine-grained and mutable nature of the target emotion. Existing studies mainly focus on achieving emotion real-time fit, while the issue of smooth transition remains understudied, affecting the overall emotional coherence of the music. In this paper, we propose REMAST to address this trade-off. Specifically, we recognize the last timestep's music emotion and fuse it with the current timestep's input emotion. The fused emotion then guides REMAST to generate the music based on the input melody. To adjust music similarity and emotion real-time fit flexibly, we downsample the original melody and feed it into the generation model. Furthermore, we design four music theory features by domain knowledge to enhance emotion information and employ semi-supervised learning to mitigate the subjective bias introduced by manual dataset annotation. According to the evaluation results, REMAST surpasses the state-of-the-art methods in objective and subjective metrics. These results demonstrate that REMAST achieves real-time fit and smooth transition simultaneously, enhancing the coherence of the generated music.

cs.SD↗

arXiv2024

The Janus Interface: How Fine-Tuning in Large Language Models Amplifies the Privacy Risks

The rapid advancements of large language models (LLMs) have raised public concerns about the privacy leakage of personally identifiable information (PII) within their extensive training datasets. Recent studies have demonstrated that an adversary could extract highly sensitive privacy data from the training data of LLMs with carefully designed prompts. However, these attacks suffer from the model's tendency to hallucinate and catastrophic forgetting (CF) in the pre-training stage, rendering the veracity of divulged PIIs negligible. In our research, we propose a novel attack, Janus, which exploits the fine-tuning interface to recover forgotten PIIs from the pre-training data in LLMs. We formalize the privacy leakage problem in LLMs and explain why forgotten PIIs can be recovered through empirical analysis on open-source language models. Based upon these insights, we evaluate the performance of Janus on both open-source language models and two latest LLMs, i.e., GPT-3.5-Turbo and LLaMA-2-7b. Our experiment results show that Janus amplifies the privacy risks by over 10 times in comparison with the baseline and significantly outperforms the state-of-the-art privacy extraction attacks including prefix attacks and in-context learning (ICL). Furthermore, our analysis validates that existing fine-tuning APIs provided by OpenAI and Azure AI Studio are susceptible to our Janus attack, allowing an adversary to conduct such an attack at a low cost.

cs.CR↗

arXiv2024

Transforming and Combining Rewards for Aligning Large Language Models

A common approach for aligning language models to human preferences is to first learn a reward model from preference data, and then use this reward model to update the language model. We study two closely related problems that arise in this approach. First, any monotone transformation of the reward model preserves preference ranking; is there a choice that is ``better'' than others? Second, we often wish to align language models to multiple properties: how should we combine multiple reward models? Using a probabilistic interpretation of the alignment procedure, we identify a natural choice for transformation for (the common case of) rewards learned from Bradley-Terry preference models. The derived transformation is straightforward: we apply a log-sigmoid function to the centered rewards, a method we term ``LSC-transformation'' (log-sigmoid-centered transformation). This transformation has two important properties. First, it emphasizes improving poorly-performing outputs, rather than outputs that already score well. This mitigates both underfitting (where some prompts are not improved) and reward hacking (where the model learns to exploit misspecification of the reward model). Second, it enables principled aggregation of rewards by linking summation to logical conjunction: the sum of transformed rewards corresponds to the probability that the output is ``good'' in all measured properties, in a sense we make precise. Experiments aligning language models to be both helpful and harmless using RLHF show substantial improvements over the baseline (non-transformed) approach.

cs.CL↗

arXiv2024

SSL-Cleanse: Trojan Detection and Mitigation in Self-Supervised Learning

Self-supervised learning (SSL) is a prevalent approach for encoding data representations. Using a pre-trained SSL image encoder and subsequently training a downstream classifier, impressive performance can be achieved on various tasks with very little labeled data. The growing adoption of SSL has led to an increase in security research on SSL encoders and associated Trojan attacks. Trojan attacks embedded in SSL encoders can operate covertly, spreading across multiple users and devices. The presence of backdoor behavior in Trojaned encoders can inadvertently be inherited by downstream classifiers, making it even more difficult to detect and mitigate the threat. Although current Trojan detection methods in supervised learning can potentially safeguard SSL downstream classifiers, identifying and addressing triggers in the SSL encoder before its widespread dissemination is a challenging task. This challenge arises because downstream tasks might be unknown, dataset labels may be unavailable, and the original unlabeled training dataset might be inaccessible during Trojan detection in SSL encoders. We introduce SSL-Cleanse as a solution to identify and mitigate backdoor threats in SSL encoders. We evaluated SSL-Cleanse on various datasets using 1200 encoders, achieving an average detection success rate of 82.2% on ImageNet-100. After mitigating backdoors, on average, backdoored encoders achieve 0.3% attack success rate without great accuracy loss, proving the effectiveness of SSL-Cleanse. The source code of SSL-Cleanse is available at https://github.com/UCF-ML-Research/SSL-Cleanse.

cs.CR↗

arXiv2024

MuDiT & MuSiT: Alignment with Colloquial Expression in Description-to-Song Generation

Amid the rising intersection of generative AI and human artistic processes, this study probes the critical yet less-explored terrain of alignment in human-centric automatic song composition. We propose a novel task of Colloquial Description-to-Song Generation, which focuses on aligning the generated content with colloquial human expressions. This task is aimed at bridging the gap between colloquial language understanding and auditory expression within an AI model, with the ultimate goal of creating songs that accurately satisfy human auditory expectations and structurally align with musical norms. Current datasets are limited due to their narrow descriptive scope, semantic gaps and inaccuracies. To overcome data scarcity in this domain, we present the Caichong Music Dataset (CaiMD). CaiMD is manually annotated by both professional musicians and amateurs, offering diverse perspectives and a comprehensive understanding of colloquial descriptions. Unlike existing datasets pre-set with expert annotations or auto-generated ones with inherent biases, CaiMD caters more sufficiently to our purpose of aligning AI-generated music with widespread user-desired results. Moreover, we propose an innovative single-stage framework called MuDiT/MuSiT for enabling effective human-machine alignment in song creation. This framework not only achieves cross-modal comprehension between colloquial language and auditory music perceptions but also ensures generated songs align with user-desired results. MuDiT/MuSiT employs one DiT/SiT model for end-to-end generation of musical components like melody, harmony, rhythm, vocals, and instrumentation. The approach ensures harmonious sonic cohesiveness amongst all generated musical components, facilitating better resonance with human auditory expectations.

cs.SD↗

arXiv2024

Describe, Explain, Plan and Select: Interactive Planning with Large Language Models Enables Open-World Multi-Task Agents

We investigate the challenge of task planning for multi-task embodied agents in open-world environments. Two main difficulties are identified: 1) executing plans in an open-world environment (e.g., Minecraft) necessitates accurate and multi-step reasoning due to the long-term nature of tasks, and 2) as vanilla planners do not consider how easy the current agent can achieve a given sub-task when ordering parallel sub-goals within a complicated plan, the resulting plan could be inefficient or even infeasible. To this end, we propose "$\underline{D}$escribe, $\underline{E}$xplain, $\underline{P}$lan and $\underline{S}$elect" ($\textbf{DEPS}$), an interactive planning approach based on Large Language Models (LLMs). DEPS facilitates better error correction on initial LLM-generated $\textit{plan}$ by integrating $\textit{description}$ of the plan execution process and providing self-$\textit{explanation}$ of feedback when encountering failures during the extended planning phases. Furthermore, it includes a goal $\textit{selector}$, which is a trainable module that ranks parallel candidate sub-goals based on the estimated steps of completion, consequently refining the initial plan. Our experiments mark the milestone of the first zero-shot multi-task agent that can robustly accomplish 70+ Minecraft tasks and nearly double the overall performances. Further testing reveals our method's general effectiveness in popularly adopted non-open-ended domains as well (i.e., ALFWorld and tabletop manipulation). The ablation and exploratory studies detail how our design beats the counterparts and provide a promising update on the $\texttt{ObtainDiamond}$ grand challenge with our approach. The code is released at https://github.com/CraftJarvis/MC-Planner.

cs.AI↗

arXiv2024

RAM: Towards an Ever-Improving Memory System by Learning from Communications

We introduce an innovative RAG-based framework with an ever-improving memory. Inspired by humans'pedagogical process, RAM utilizes recursively reasoning-based retrieval and experience reflections to continually update the memory and learn from users' communicative feedback, namely communicative learning. Extensive experiments with both simulated and real users demonstrate significant improvements over traditional RAG and self-knowledge methods, particularly excelling in handling false premise and multi-hop questions. Furthermore, RAM exhibits promising adaptability to various feedback and retrieval methods, showcasing its potential for advancing AI capabilities in dynamic knowledge acquisition and lifelong learning.

cs.AI↗

arXiv2024

Layer-Dependent Charge State Lifetime of Single Se Vacancies in WSe$_2$

Defect engineering in two-dimensional semiconductors has been exploited to tune the optoelectronic properties and introduce new quantum states in the band gap. Chalcogen vacancies in transition metal dichalcogenides in particular have been found to strongly impact charge carrier concentration and mobility in 2D transistors as well as feature sub-gap emission and single-photon response. In this letter, we investigate the layer-dependent charge state lifetime of Se vacancies in WSe$_2$. In one monolayer WSe$_2$, we observe ultrafast charge transfer from the lowest unoccupied orbital of the top Se vacancy to the graphene substrate within (1.0 $\pm$ 0.2) ps measured via the current saturation in scanning tunneling approach curves. For Se vacancies decoupled by TMD multilayers, we find a sub-exponential increase of the charge lifetime from (62 $\pm$ 14) ps in bilayer to few nanoseconds in four-layer WSe$_2$, alongside a reduction of the defect state binding energy. Additionally, we attribute the continuous suppression and energy shift of the dI/dV in-gap defect state resonances at very close tip--sample distances to a current saturation effect. Our results provide a key measure of the layer-dependent charge transfer rate of chalcogen vacancies in TMDs.

cond-mat.mes-hall↗

arXiv2024

Privacy-Preserved Neural Graph Databases

In the era of large language models (LLMs), efficient and accurate data retrieval has become increasingly crucial for the use of domain-specific or private data in the retrieval augmented generation (RAG). Neural graph databases (NGDBs) have emerged as a powerful paradigm that combines the strengths of graph databases (GDBs) and neural networks to enable efficient storage, retrieval, and analysis of graph-structured data which can be adaptively trained with LLMs. The usage of neural embedding storage and Complex neural logical Query Answering (CQA) provides NGDBs with generalization ability. When the graph is incomplete, by extracting latent patterns and representations, neural graph databases can fill gaps in the graph structure, revealing hidden relationships and enabling accurate query answering. Nevertheless, this capability comes with inherent trade-offs, as it introduces additional privacy risks to the domain-specific or private databases. Malicious attackers can infer more sensitive information in the database using well-designed queries such as from the answer sets of where Turing Award winners born before 1950 and after 1940 lived, the living places of Turing Award winner Hinton are probably exposed, although the living places may have been deleted in the training stage due to the privacy concerns. In this work, we propose a privacy-preserved neural graph database (P-NGDB) framework to alleviate the risks of privacy leakage in NGDBs. We introduce adversarial training techniques in the training stage to enforce the NGDBs to generate indistinguishable answers when queried with private information, enhancing the difficulty of inferring sensitive information through combinations of multiple innocuous queries.

cs.DB↗

arXiv2024

MuChin: A Chinese Colloquial Description Benchmark for Evaluating Language Models in the Field of Music

The rapidly evolving multimodal Large Language Models (LLMs) urgently require new benchmarks to uniformly evaluate their performance on understanding and textually describing music. However, due to semantic gaps between Music Information Retrieval (MIR) algorithms and human understanding, discrepancies between professionals and the public, and low precision of annotations, existing music description datasets cannot serve as benchmarks. To this end, we present MuChin, the first open-source music description benchmark in Chinese colloquial language, designed to evaluate the performance of multimodal LLMs in understanding and describing music. We established the Caichong Music Annotation Platform (CaiMAP) that employs an innovative multi-person, multi-stage assurance method, and recruited both amateurs and professionals to ensure the precision of annotations and alignment with popular semantics. Utilizing this method, we built a dataset with multi-dimensional, high-precision music annotations, the Caichong Music Dataset (CaiMD), and carefully selected 1,000 high-quality entries to serve as the test set for MuChin. Based on MuChin, we analyzed the discrepancies between professionals and amateurs in terms of music description, and empirically demonstrated the effectiveness of annotated data for fine-tuning LLMs. Ultimately, we employed MuChin to evaluate existing music understanding models on their ability to provide colloquial descriptions of music. All data related to the benchmark, along with the scoring code and detailed appendices, have been open-sourced (https://github.com/CarlWangChina/MuChin/).

cs.SD↗

arXiv2024

CLoG: Benchmarking Continual Learning of Image Generation Models

Continual Learning (CL) poses a significant challenge in Artificial Intelligence, aiming to mirror the human ability to incrementally acquire knowledge and skills. While extensive research has focused on CL within the context of classification tasks, the advent of increasingly powerful generative models necessitates the exploration of Continual Learning of Generative models (CLoG). This paper advocates for shifting the research focus from classification-based CL to CLoG. We systematically identify the unique challenges presented by CLoG compared to traditional classification-based CL. We adapt three types of existing CL methodologies, replay-based, regularization-based, and parameter-isolation-based methods to generative tasks and introduce comprehensive benchmarks for CLoG that feature great diversity and broad task coverage. Our benchmarks and results yield intriguing insights that can be valuable for developing future CLoG methods. Additionally, we will release a codebase designed to facilitate easy benchmarking and experimentation in CLoG publicly at https://github.com/linhaowei1/CLoG. We believe that shifting the research focus to CLoG will benefit the continual learning community and illuminate the path for next-generation AI-generated content (AIGC) in a lifelong learning paradigm.

cs.LG↗

arXiv2024

Texture-guided Coding for Deep Features

With the rapid development of machine vision technology in recent years, many researchers have begun to focus on feature compression that is better suited for machine vision tasks. The target of feature compression is deep features, which arise from convolution in the middle layer of a pre-trained convolutional neural network. However, due to the large volume of data and high level of abstraction of deep features, their application is primarily limited to machine-centric scenarios, which poses significant constraints in situations requiring human-computer interaction. This paper investigates features and textures and proposes a texture-guided feature compression strategy based on their characteristics. Specifically, the strategy comprises feature layers and texture layers. The feature layers serve the machine, including a feature selection module and a feature reconstruction network. With the assistance of texture images, they selectively compress and transmit channels relevant to visual tasks, reducing feature data while providing high-quality features for the machine. The texture layers primarily serve humans and consist of an image reconstruction network. This image reconstruction network leverages features and texture images to reconstruct preview images for humans. Our method fully exploits the characteristics of texture and features. It eliminates feature redundancy, reconstructs high-quality preview images for humans, and supports decision-making. The experimental results demonstrate excellent performance when employing our proposed method to compress the deep features.

cs.CV↗