arXiv subjects

Yue Wu

Publications and source records attributed to Yue Wu.

At least 181 records · Page 10Linked to original sources

arXiv2024

Phi-4 Technical Report

We present phi-4, a 14-billion parameter language model developed with a training recipe that is centrally focused on data quality. Unlike most language models, where pre-training is based primarily on organic data sources such as web content or code, phi-4 strategically incorporates synthetic data throughout the training process. While previous models in the Phi family largely distill the capabilities of a teacher model (specifically GPT-4), phi-4 substantially surpasses its teacher model on STEM-focused QA capabilities, giving evidence that our data-generation and post-training techniques go beyond distillation. Despite minimal changes to the phi-3 architecture, phi-4 achieves strong performance relative to its size -- especially on reasoning-focused benchmarks -- due to improved data, training curriculum, and innovations in the post-training scheme.

cs.CL↗

arXiv2024

LINKs: Large Language Model Integrated Management for 6G Empowered Digital Twin NetworKs

In the rapidly evolving landscape of digital twins (DT) and 6G networks, the integration of large language models (LLMs) presents a novel approach to network management. This paper explores the application of LLMs in managing 6G-empowered DT networks, with a focus on optimizing data retrieval and communication efficiency in smart city scenarios. The proposed framework leverages LLMs for intelligent DT problem analysis and radio resource management (RRM) in fully autonomous way without any manual intervention. Our proposed framework -- LINKs, builds up a lazy loading strategy which can minimize transmission delay by selectively retrieving the relevant data. Based on the data retrieval plan, LLMs transform the retrieval task into an numerical optimization problem and utilizing solvers to build an optimal RRM, ensuring efficient communication across the network. Simulation results demonstrate the performance improvements in data planning and network management, highlighting the potential of LLMs to enhance the integration of DT and 6G technologies.

cs.NI↗

arXiv2024

The random periodic solutions for McKean-Vlasov stochastic differential equations

In this paper, we study well-posedness of random periodic solutions of stochastic differential equations (SDEs) of McKean-Vlasov type driven by a two-sided Brownian motion, where the random periodic behaviour is characterised by the equations' long-time behaviour. Given the well-known connection between McKean-Vlasov SDEs and interacting particle systems, we show propagation of chaos and that the key properties of the interacting particle systems recover those of the McKean-Vlasov SDEs in the particle limit. All results in the present work are shown under two settings: fully and partially dissipative case. Each setting has its challenges and limitations. For instance, weakening full dissipativity to partial dissipativity demands stronger structural assumptions on the equations' dynamics and yields random periodic behaviour in the weak sense instead of pathwise sense (as in the full dissipativity case). The proof mechanisms are close but fundamentally different.

math.PR↗

arXiv2024

Electrical contact with dielectric breakdown of interfacial gap

Electrical contact is fundamental to almost every aspect of modern industry, including the fast-growing electric vehicle industry. In metallic contacts in atmospheric conditions, most of the electrical current passes via the micro-junctions formed between two electrodes. The classic electrical contact theory predicts an infinite current density at the circular contact periphery. In the present work, we explore the influence of the dielectric breakdown of air outside the contact area on the electrical contact interface. Incorporating the discharging boundary condition governed by the modified Paschen law, we develop the numerical model as well as two sets of closed-form solutions for low applied voltage cases where two electrodes are in solid-solid contact and complete separation, respectively. For Hertzian contact, the present work theoretically proves that the ignorance of discharge can lead to a singular current density at the contact periphery and an overestimation of the electrical contact resistance. The current density monotonically increases along the radial direction to a finite value at the contact area periphery, followed by a monotonic drop within the discharge zone. The present study serves as a foundation for the modeling of discharging rough surface electrical contact and sheds light on the machine element surface damages caused by the electrical discharge machining.

cond-mat.soft↗

arXiv2024

A projected Euler Method for Random Periodic Solutions of Semi-linear SDEs with non-globally Lipschitz coefficients

The present work introduces and investigates an explicit time discretization scheme, called the projected Euler method,to numerically approximate random periodic solutions of semi-linear SDEs under non-globally Lipschitz conditions. The existence of the random periodic solution is demonstrated as the limit of the pull-back of the discretized SDE. Without relying on a priori high-order moment bounds of the numerical approximations, the mean square convergence rate of the approximation scheme is proved to be order $0.5$ for SDEs with multiplicative noise and order $1$ for SDEs with additive noise. Numerical examples are also provided to validate our theoretical findings.

math.NA↗

arXiv2024

Delving into the Reversal Curse: How Far Can Large Language Models Generalize?

While large language models (LLMs) showcase unprecedented capabilities, they also exhibit certain inherent limitations when facing seemingly trivial tasks. A prime example is the recently debated "reversal curse", which surfaces when models, having been trained on the fact "A is B", struggle to generalize this knowledge to infer that "B is A". In this paper, we examine the manifestation of the reversal curse across various tasks and delve into both the generalization abilities and the problem-solving mechanisms of LLMs. This investigation leads to a series of significant insights: (1) LLMs are able to generalize to "B is A" when both A and B are presented in the context as in the case of a multiple-choice question. (2) This generalization ability is highly correlated to the structure of the fact "A is B" in the training documents. For example, this generalization only applies to biographies structured in "[Name] is [Description]" but not to "[Description] is [Name]". (3) We propose and verify the hypothesis that LLMs possess an inherent bias in fact recalling during knowledge application, which explains and underscores the importance of the document structure to successful learning. (4) The negative impact of this bias on the downstream performance of LLMs can hardly be mitigated through training alone. These findings offer a novel perspective on interpreting LLMs' generalization through their intrinsic mechanisms and provide insights for developing more effective learning methods. Our code and data are available at https://github.com/alibaba/thinking_bias.git.

cs.CL↗

arXiv2024

Distribution of plastics of various sizes and densities in the global ocean from a 3D Eulerian model

We develop a 3D Eulerian model to study the transport and distribution of microplastics in the global ocean. Among other benefits that will be discussed in the paper, one unique feature of our model is that it takes into consideration the effect of properties of particles (size and density, the former for the first time) to their vertical terminal velocity. With ocean current velocity taken from ECCOv4r4, a dataset generated from a data-assimilated MITgcm reanalysis, our model is integrated for 26 years for particles of different properties with their stationary patterns studied. We find that only low-density particles with sufficient size (e.g. density $900kg/m^3$ with size $\gtrsim 10 μm$) aggregate in the five subtropical gyres observed in previous studies. In contrast, particles of smaller size ($\sim 1 μm$), irrespective of their density, behave like neutrally buoyant particles with a weaker pattern on the surface and a deeper penetration into depth (up to about 1km deep). In addition, we observe seasonal variations of floating particle concentration on the ocean surface, which reasonably agree with the satellite observation by Cyclone Global Navigation Satellite System (CYGNSS) in terms of the phase of the variation. We find that the seasonal variation of the surface particle concentration correlates well with the variation of the mixing layer (ML) depth globally, due to an almost uniform vertical distribution of particles in the ML with total amount of particles conserved.

physics.ao-ph↗

arXiv2024

Hybrid Precoding with Per-Beam Timing Advance for Asynchronous Cell-free mmWave Massive MIMO-OFDM Systems

Cell-free massive multiple-input-multiple-output (CF-mMIMO) is regarded as one of the promising technologies for next-generation wireless networks. However, due to its distributed architecture, geographically separated access points (APs) jointly serve a large number of user-equipments (UEs), there will inevitably be a discrepancies in the arrival time of transmitted signals. In this paper, we investigate millimeter-wave (mmWave) CF-mMIMO orthogonal frequency division multiplexing (OFDM) systems with asynchronous reception in a wide area coverage scenario, where asynchronous timing offsets may extend far beyond the cyclic prefix (CP) range. A comprehensive asynchronous beam-domain signal transmission model is presented for mmWave CF-mMIMO-OFDM systems in both downlink and uplink, incorporating phase offset, inter-carrier interference (ICI) and inter-symbol interference (ISI). To address the issue of asynchronous reception, we propose a novel per-beam timing advance (PBTA) hybrid precoding architecture and analyze the spectral efficiency (SE) in the beam domain for downlink and uplink asynchronous receptions. Both scalable centralized and distributed implementations are taken into account, and the asynchronous delay phase is utilized to design precoding/combining vectors. Furthermore, we formulate the sum rate maximization problem and develop two low-complexity joint beam selection and UE association algorithms considering the impact of asynchronous timing offset exceeding the CP range. Simulation results demonstrate that the performance will be severely limited by ICI and ISI, and our proposed PBTA hybrid precoding architecture effectively mitigates asynchronous interference compared to the nearest AAU/UE-based timing-advance scheme. Additionally, numerical results show that our proposed low-complexity joint beam selection and UE association algorithms achieve superior SE performance.

eess.SP↗

arXiv2024

Enhancing LLM's Cognition via Structurization

When reading long-form text, human cognition is complex and structurized. While large language models (LLMs) process input contexts through a causal and sequential perspective, this approach can potentially limit their ability to handle intricate and complex inputs effectively. To enhance LLM's cognition capability, this paper presents a novel concept of context structurization. Specifically, we transform the plain, unordered contextual sentences into well-ordered and hierarchically structurized elements. By doing so, LLMs can better grasp intricate and extended contexts through precise attention and information-seeking along the organized structures. Extensive evaluations are conducted across various model architectures and sizes (including a series of auto-regressive LLMs as well as BERT-like masking models) on a diverse set of NLP tasks (e.g., context-based question-answering, exhaustive hallucination evaluation, and passage-level dense retrieval). Empirical results show consistent and significant performance gains afforded by a single-round structurization. In particular, we boost the open-sourced LLaMA2-70B model to achieve comparable performance against GPT-3.5-Turbo as the hallucination evaluator. Besides, we show the feasibility of distilling advanced LLMs' language processing abilities to a smaller yet effective StruXGPT-7B to execute structurization, addressing the practicality of our approach. Code is available at https://github.com/alibaba/struxgpt.

cs.CL↗

arXiv2024

PrivacyMind: Large Language Models Can Be Contextual Privacy Protection Learners

The proliferation of Large Language Models (LLMs) has driven considerable interest in fine-tuning them with domain-specific data to create specialized language models. Nevertheless, such domain-specific fine-tuning data often contains contextually sensitive personally identifiable information (PII). Direct fine-tuning of LLMs on this data without privacy protection poses a risk of data leakage of sensitive PII during inference time. To address this challenge, we introduce Contextual Privacy Protection Language Models (PrivacyMind), a novel paradigm for fine-tuning LLMs that effectively injects domain-specific knowledge while safeguarding inference-time data privacy. Our work offers a theoretical analysis for model design and benchmarks various techniques such as corpus curation, penalty-based unlikelihood in training loss, instruction-based tuning, etc. Extensive experiments across diverse datasets and scenarios demonstrate the effectiveness of our approaches. In particular, instruction tuning with both positive and negative examples stands out as a promising method, effectively protecting private data while enhancing the model's knowledge. Our work underscores the potential for Large Language Models as robust contextual privacy protection learners. The complete code and data for the work can be found at https://github.com/Yijia-Xiao/PrivacyMind.

cs.CL↗

arXiv2024

INSIDE: LLMs' Internal States Retain the Power of Hallucination Detection

Knowledge hallucination have raised widespread concerns for the security and reliability of deployed LLMs. Previous efforts in detecting hallucinations have been employed at logit-level uncertainty estimation or language-level self-consistency evaluation, where the semantic information is inevitably lost during the token-decoding procedure. Thus, we propose to explore the dense semantic information retained within LLMs' \textbf{IN}ternal \textbf{S}tates for halluc\textbf{I}nation \textbf{DE}tection (\textbf{INSIDE}). In particular, a simple yet effective \textbf{EigenScore} metric is proposed to better evaluate responses' self-consistency, which exploits the eigenvalues of responses' covariance matrix to measure the semantic consistency/diversity in the dense embedding space. Furthermore, from the perspective of self-consistent hallucination detection, a test time feature clipping approach is explored to truncate extreme activations in the internal states, which reduces overconfident generations and potentially benefits the detection of overconfident hallucinations. Extensive experiments and ablation studies are performed on several popular LLMs and question-answering (QA) benchmarks, showing the effectiveness of our proposal.

cs.CL↗

arXiv2024

Variance-Aware Regret Bounds for Stochastic Contextual Dueling Bandits

Dueling bandits is a prominent framework for decision-making involving preferential feedback, a valuable feature that fits various applications involving human interaction, such as ranking, information retrieval, and recommendation systems. While substantial efforts have been made to minimize the cumulative regret in dueling bandits, a notable gap in the current research is the absence of regret bounds that account for the inherent uncertainty in pairwise comparisons between the dueling arms. Intuitively, greater uncertainty suggests a higher level of difficulty in the problem. To bridge this gap, this paper studies the problem of contextual dueling bandits, where the binary comparison of dueling arms is generated from a generalized linear model (GLM). We propose a new SupLinUCB-type algorithm that enjoys computational efficiency and a variance-aware regret bound $\tilde O\big(d\sqrt{\sum_{t=1}^Tσ_t^2} + d\big)$, where $σ_t$ is the variance of the pairwise comparison in round $t$, $d$ is the dimension of the context vectors, and $T$ is the time horizon. Our regret bound naturally aligns with the intuitive expectation in scenarios where the comparison is deterministic, the algorithm only suffers from an $\tilde O(d)$ regret. We perform empirical experiments on synthetic data to confirm the advantage of our method over previous variance-agnostic algorithms.

cs.LG↗

arXiv2024

Triple Point Masking

Existing 3D mask learning methods encounter performance bottlenecks under limited data, and our objective is to overcome this limitation. In this paper, we introduce a triple point masking scheme, named TPM, which serves as a scalable framework for pre-training of masked autoencoders to achieve multi-mask learning for 3D point clouds. Specifically, we augment the baselines with two additional mask choices (i.e., medium mask and low mask) as our core insight is that the recovery process of an object can manifest in diverse ways. Previous high-masking schemes focus on capturing the global representation but lack the fine-grained recovery capability, so that the generated pre-trained weights tend to play a limited role in the fine-tuning process. With the support of the proposed TPM, available methods can exhibit more flexible and accurate completion capabilities, enabling the potential autoencoder in the pre-training stage to consider multiple representations of a single 3D object. In addition, an SVM-guided weight selection module is proposed to fill the encoder parameters for downstream networks with the optimal weight during the fine-tuning stage, maximizing linear accuracy and facilitating the acquisition of intricate representations for new objects. Extensive experiments show that the four baselines equipped with the proposed TPM achieve comprehensive performance improvements on various downstream tasks. Our code and models are available at https://github.com/liujia99/TPM.

cs.CV↗

arXiv2024

LEGO: QEC Decoding System Architecture for Dynamic Circuits

Quantum error correction (QEC) is a critical component of FTQC; the QEC decoder is an important part of Classical Computing for Quantum or C4Q. Recent years have seen fast development in real-time QEC decoders. Existing efforts to build real-time decoders have yet to achieve a critical milestone: decoding dynamic logical circuits with error-corrected readout and feed forward. Achieving this requires significant engineering effort to adapt and reconfigure the decoders during runtime, depending on the branching of the logical circuit. We present a QEC decoder architecture called LEGO, with the ambitious goal of supporting dynamic logical operations. LEGO employs a novel abstraction called the decoding block to describe the decoding problem of a dynamic logical circuit. Moreover, decoding blocks can be combined with three other ideas to improve the efficiency, accuracy and latency of the decoder. First, they provide data and task parallelisms when combined with fusion-based decoding. Second, they can exploit the pipeline parallelism inside multi-stage decoders. Finally, they serve as basic units of work for computational resource management. Using decoding blocks, LEGO can be easily reconfigured to support all QEC settings and to easily accommodate innovations in three interdependent fields: code, logical operations and qubit hardware. In contrast, existing decoders are highly specialized to a specific QEC setting, which leads to redundant research and engineering efforts, slows down innovation, and further fragments the nascent quantum computing industry.

quant-ph↗

arXiv2024

Self-Play Preference Optimization for Language Model Alignment

Standard reinforcement learning from human feedback (RLHF) approaches relying on parametric models like the Bradley-Terry model fall short in capturing the intransitivity and irrationality in human preferences. Recent advancements suggest that directly working with preference probabilities can yield a more accurate reflection of human preferences, enabling more flexible and accurate language model alignment. In this paper, we propose a self-play-based method for language model alignment, which treats the problem as a constant-sum two-player game aimed at identifying the Nash equilibrium policy. Our approach, dubbed Self-Play Preference Optimization (SPPO), utilizes iterative policy updates to provably approximate the Nash equilibrium. Additionally, we propose a new SPPO objective which is both strongly motivated by theory and is simple and effective in practice. In our experiments, using only 60k prompts (without responses) from the UltraFeedback dataset and without any prompt augmentation, by leveraging a pre-trained preference model PairRM with only 0.4B parameters, SPPO can obtain a model from fine-tuning Mistral-7B-Instruct-v0.2 that achieves the state-of-the-art length-controlled win-rate of 28.53% against GPT-4-Turbo on AlpacaEval 2.0. It also outperforms the (iterative) DPO and IPO on MT-Bench, Arena-Hard, and the Open LLM Leaderboard. Starting from a stronger base model Llama-3-8B-Instruct, we are able to achieve a length-controlled win rate of 38.77%. Notably, the strong performance of SPPO is achieved without additional external supervision (e.g., responses, preferences, etc.) from GPT-4 or other stronger language models. Codes are available at https://github.com/uclaml/SPPO.

cs.LG↗

arXiv2024

Llama SLayer 8B: Shallow Layers Hold the Key to Knowledge Injection

As a manner to augment pre-trained large language models (LLM), knowledge injection is critical to develop vertical domain large models and has been widely studied. Although most current approaches, including parameter-efficient fine-tuning (PEFT) and block expansion methods, uniformly apply knowledge across all LLM layers, it raises the question: are all layers equally crucial for knowledge injection? We begin by evaluating the importance of each layer in finding the optimal layer range for knowledge injection. Intuitively, the more important layers should play a more critical role in knowledge injection and deserve a denser injection. We observe performance dips in question-answering benchmarks after the removal or expansion of the shallow layers, and the degradation shrinks as the layer gets deeper, indicating that the shallow layers hold the key to knowledge injection. This insight leads us to propose the S strategy, a post-pretraining strategy of selectively enhancing shallow layers while pruning the less effective deep ones. Based on this strategy, we introduce Llama Slayer-8B and Llama Slayer-8B-Instruct. We experimented on the corpus of code $\&$ math and demonstrated the effectiveness of our strategy. Further experiments across different LLM, Mistral-7B, and a legal corpus confirmed the general applicability of the approach, underscoring its wide-ranging efficacy. Our code is available at: \https://github.com/txchen-USTC/Llama-Slayer

cs.CL↗

arXiv2024

FPGA-based Distributed Union-Find Decoder for Surface Codes

A fault-tolerant quantum computer must decode and correct errors faster than they appear to prevent exponential slowdown due to error correction. The Union-Find (UF) decoder is promising with an average time complexity slightly higher than $O(d^3)$. We report a distributed version of the UF decoder that exploits parallel computing resources for further speedup. Using an FPGA-based implementation, we empirically show that this distributed UF decoder has a sublinear average time complexity with regard to $d$, given $O(d^3)$ parallel computing resources. The decoding time per measurement round decreases as $d$ increases, the first time for a quantum error decoder. The implementation employs a scalable architecture called Helios that organizes parallel computing resources into a hybrid tree-grid structure. Using a Xilinx VCU129 FPGA, we successfully implement $d$ up to 21 with an average decoding time of 11.5 ns per measurement round under 0.1\% phenomenological noise, and 23.7 ns for $d=17$ under equivalent circuit-level noise. This performance is significantly faster than any existing decoder implementation. Furthermore, we show that Helios can optimize for resource efficiency by decoding $d=51$ on a Xilinx VCU129 FPGA with an average latency of 544ns per measurement round.

quant-ph↗

arXiv2024

Protein Conformation Generation via Force-Guided SE(3) Diffusion Models

The conformational landscape of proteins is crucial to understanding their functionality in complex biological processes. Traditional physics-based computational methods, such as molecular dynamics (MD) simulations, suffer from rare event sampling and long equilibration time problems, hindering their applications in general protein systems. Recently, deep generative modeling techniques, especially diffusion models, have been employed to generate novel protein conformations. However, existing score-based diffusion methods cannot properly incorporate important physical prior knowledge to guide the generation process, causing large deviations in the sampled protein conformations from the equilibrium distribution. In this paper, to overcome these limitations, we propose a force-guided SE(3) diffusion model, ConfDiff, for protein conformation generation. By incorporating a force-guided network with a mixture of data-based score models, ConfDiff can generate protein conformations with rich diversity while preserving high fidelity. Experiments on a variety of protein conformation prediction tasks, including 12 fast-folding proteins and the Bovine Pancreatic Trypsin Inhibitor (BPTI), demonstrate that our method surpasses the state-of-the-art method.

q-bio.BM↗