arXiv subjects

Zheng Liu

Publications and source records attributed to Zheng Liu.

At least 181 records · Page 10Linked to original sources

arXiv2025

Quantum entanglement and Einstein-Podolsky-Rosen steering in ultrastrongly light-matter coupled system

This work presents a scheme for engineering quantum entanglement and Einstein-Podolsky-Rosen (EPR) steering with Gaussian measurements based on the quantum Hopfield model that incorporates a common thermal reservoir. We begin by examining quantum correlations, specifically quantum entanglement and EPR steering, in the ground state. These quantum correlations primarily stem from squeezing interactions in weak and normal strong coupling regimes. As the coupling strength increases, especially upon entering the ultrastrong coupling regime, the correlations emerge from the combined effect of squeezing and mix-mode interactions. Importantly, this scenario enables the realization of two-way EPR steering. Moreover, lower optical frequencies enhance both quantum entanglement and EPR steering. Further, when considering thermal effects, the ultrastrong and deep strong coupling regimes, paired with lower optical frequencies, lead to improved entanglement. The one-way EPR steering for resonant case can be effectively controlled in the ultrastrong and deep strong coupling regimes which originates from the asymmetry of subsystem and reservoir coupling induced by the diamagnetic term. Additionally, one-way EPR steering can also be produced for nonresonant case. In this case, the asymmetry of the subsystem and reservoir originates from the combined effect of nonresonant frequencies and diamagnetic term. Our findings have the potential to inspire further research into quantum information processing that leverages light-matter entanglement and EPR steering.

quant-ph↗

arXiv2025

Singular transport in non-equilibrium strongly internal-coupled 1D tilted field spin-1/2 chain

Non-equilibrium spin-chain systems have been attracting increasing interest in energy transport. This work studies a one-dimensional non-equilibrium Ising chain immersed in a tilted magnetic field, every spin contacts a Boson reservoir with the dissipative system-environment interaction. We analytically investigate the dynamics and the steady-state energy transport taking advantage of the Born-Markov-secular master equation. In the longitudinal field, one can find that the non dissipative $N^\prime$ spins decompose the spin chain into $N^\prime +1$ independent subchains and block the heat currents from the hot end to the cool end. Moreover, for the non-dissipative $μ$th spin, its nearest two bulk spins become the nodal spins in the subchains and have the corresponding energy correction of $\pm J_{μ-1,μ}$ and $\pm J_{μ,μ+1}$ depending on the excited/ground state of the $μ$th spin. Therefore, a magnetically controlled heat modulator can be designed by adjusting the direction of the magnetic field in which the non-dissipative spin is located. For the transverse field case, the whole Hilbert space of the chain can always be divided into two independent subspaces regardless of whether the bulk spin is dissipative. This work provides new insight into the dynamics and energy transport of the dissipative Ising model.

quant-ph↗

arXiv2025

Video-XL-Pro: Reconstructive Token Compression for Extremely Long Video Understanding

Despite advanced token compression techniques, existing multimodal large language models (MLLMs) still struggle with hour-long video understanding. In this work, we propose Video-XL-Pro, an efficient method for extremely long video understanding, built upon Reconstructive Compression of Tokens (ReCoT), a learnable module that leverages self-supervised learning to generate comprehensive and compact video tokens. ReCoT introduces two key components: (i) Dynamic Token Synthesizer (DTS): DTS generates pseudo-video tokens from static image tokens by learning intra-token relationships, which are then used in masked video modeling. (ii) Semantic-Guided Masking (SGM): SGM adaptively masks redundant visual tokens to facilitate more effective reconstructive learning. To improve training efficiency in MLLMs fine-tuning, we introduce a video-specific dataset pruning strategy and design a simple yet Query-aware Selector that enables the model to precisely locate query-relevant video tokens. With only 3B parameters, Video-XL-Pro outperforms most 7B models trained on larger datasets across multiple long video understanding benchmarks. Moreover, it can process over 8K frames on a single A100 GPU while maintaining high-quality performance.

cs.CV↗

arXiv2025

HeRB: Heterophily-Resolved Structure Balancer for Graph Neural Networks

Recent research has witnessed the remarkable progress of Graph Neural Networks (GNNs) in the realm of graph data representation. However, GNNs still encounter the challenge of structural imbalance. Prior solutions to this problem did not take graph heterophily into account, namely that connected nodes process distinct labels or features, thus resulting in a deficiency in effectiveness. Upon verifying the impact of heterophily on solving the structural imbalance problem, we propose to rectify the heterophily first and then transfer homophilic knowledge. To the end, we devise a method named HeRB (Heterophily-Resolved Structure Balancer) for GNNs. HeRB consists of two innovative components: 1) A heterophily-lessening augmentation module which serves to reduce inter-class edges and increase intra-class edges; 2) A homophilic knowledge transfer mechanism to convey homophilic information from head nodes to tail nodes. Experimental results demonstrate that HeRB achieves superior performance on two homophilic and six heterophilic benchmark datasets, and the ablation studies further validate the efficacy of two proposed components.

cs.LG↗

arXiv2025

Flying through cluttered and dynamic environments with LiDAR

Navigating unmanned aerial vehicles (UAVs) through cluttered and dynamic environments remains a significant challenge, particularly when dealing with fast-moving or sudden-appearing obstacles. This paper introduces a complete LiDAR-based system designed to enable UAVs to avoid various moving obstacles in complex environments. Benefiting the high computational efficiency of perception and planning, the system can operate in real time using onboard computing resources with low latency. For dynamic environment perception, we have integrated our previous work, M-detector, into the system. M-detector ensures that moving objects of different sizes, colors, and types are reliably detected. For dynamic environment planning, we incorporate dynamic object predictions into the integrated planning and control (IPC) framework, namely DynIPC. This integration allows the UAV to utilize predictions about dynamic obstacles to effectively evade them. We validate our proposed system through both simulations and real-world experiments. In simulation tests, our system outperforms state-of-the-art baselines across several metrics, including success rate, time consumption, average flight time, and maximum velocity. In real-world trials, our system successfully navigates through forests, avoiding moving obstacles along its path.

cs.RO↗

arXiv2025

MemoRAG: Boosting Long Context Processing with Global Memory-Enhanced Retrieval Augmentation

Processing long contexts presents a significant challenge for large language models (LLMs). While recent advancements allow LLMs to handle much longer contexts than before (e.g., 32K or 128K tokens), it is computationally expensive and can still be insufficient for many applications. Retrieval-Augmented Generation (RAG) is considered a promising strategy to address this problem. However, conventional RAG methods face inherent limitations because of two underlying requirements: 1) explicitly stated queries, and 2) well-structured knowledge. These conditions, however, do not hold in general long-context processing tasks. In this work, we propose MemoRAG, a novel RAG framework empowered by global memory-augmented retrieval. MemoRAG features a dual-system architecture. First, it employs a light but long-range system to create a global memory of the long context. Once a task is presented, it generates draft answers, providing useful clues for the retrieval tools to locate relevant information within the long context. Second, it leverages an expensive but expressive system, which generates the final answer based on the retrieved information. Building upon this fundamental framework, we realize the memory module in the form of KV compression, and reinforce its memorization and cluing capacity from the Generation quality's Feedback (a.k.a. RLGF). In our experiments, MemoRAG achieves superior performances across a variety of long-context evaluation tasks, not only complex scenarios where traditional RAG methods struggle, but also simpler ones where RAG is typically applied.

cs.CL↗

arXiv2025

Fractionalized Kohn-Sham Scheme for Strongly Correlated Electrons

We propose to expand the territory of density functional theory to strongly correlated electrons by reformulating the Kohn-Sham scheme in the representation of fractionalized particles. We call it the ``KS* scheme.'' Using inhomogeneous $t$-$J$ chains as a test bed, we show that the KS* scheme with simple local density approximation is able to achieve accurate ground-state energy and density distribution comparable to the density matrix renormalization group method, while the computational complexity is much lower.

physics.comp-ph↗

arXiv2025

Resolving phonon-mediated superconducting pairing symmetries from first-principles calculation

The quest for topological superconductors triggers revived interests in resolving non-s-wave pairing channels mediated by phonons. While density functional theory and density functional perturbtaion theory have established a powerful framework to calculate electron-phonon couplings in real materials in a first-principles way, its application is largely limited to conventional s-wave superconductivity. Here, we formulate an efficient and simple-to-use algorithm for first-principles pairing channel analysis, and apply it to several representative material systems.

cond-mat.supr-con↗

arXiv2025

Boosting classical and quantum nonlinear processes in ultrathin van der Waals materials

Understanding and controlling nonlinear processes is crucial for engineering light-matter interaction and generating non-classical light. A significant challenge in ultra-thin nonlinear materials is the marked diminution of the nonlinear conversion efficiency due to the reduced light-matter interaction length and, in many cases, the centrosymmetric crystalline structures. Here we relax these limitations and report a giant boost of classical and quantum nonlinear processes in ultrathin van der Waals materials. Specifically, with a metal-nonlinear material heterostructure we enhance classical second-harmonic generation in h-BN flakes by two-orders of magnitude. Moreover, we have engineered a metal-SiO2-nonlinear material heterostructure resulting in a remarkable two orders of magnitude augmentation of the quantum spontaneous parametric down-conversion (SPDC) in NbOCl2 flakes. Notably, we demonstrate SPDC in a 16 nm-thick NbOCl2 flake integrated into the proposed structure. These findings simplify on-chip quantum state engineering and accelerate the use of van der Waals materials in nonlinear optoelectronics.

quant-ph↗

arXiv2025

Two-dimensional atomic-scale ultrathin lateral heterostructures

Ultrathin lateral heterostructures of monolayer MoS2 and WS2 have successfully been realized with the metal-organic chemical vapor deposition method. Atomic-resolution HAADF-STEM observations have revealed that the junction widths of lateral heterostructures range from several nanometers to single-atom thickness, the thinnest heterojunction in theory. The interfaces are atomically flat with minimal mixing between MoS2 and WS2, originating from rapid and abrupt switching of the source supply. Due to one-dimensional interfaces and broken rotational symmetry, the resulting ultrathin lateral heterostructures, 1~2 mixed-dimensional structures, can show emergent optical/electronic properties. The MOCVD growth developed in this work allows us to access various ultrathin lateral heterostructures, leading to future exploration of their emergent properties absent in each component alone.

cond-mat.mtrl-sci↗

arXiv2025

Unveiling the Ignorance of MLLMs: Seeing Clearly, Answering Incorrectly

Multimodal Large Language Models (MLLMs) have displayed remarkable performance in multi-modal tasks, particularly in visual comprehension. However, we reveal that MLLMs often generate incorrect answers even when they understand the visual content. To this end, we manually construct a benchmark with 12 categories and design evaluation metrics that assess the degree of error in MLLM responses even when the visual content is seemingly understood. Based on this benchmark, we test 15 leading MLLMs and analyze the distribution of attention maps and logits of some MLLMs. Our investigation identifies two primary issues: 1) most instruction tuning datasets predominantly feature questions that 'directly' relate to the visual content, leading to a bias in MLLMs' responses to other indirect questions, and 2) MLLMs' attention to visual tokens is notably lower than to system and question tokens. We further observe that attention scores between questions and visual tokens as well as the model's confidence in the answers are lower in response to misleading questions than to straightforward ones. To address the first challenge, we introduce a paired positive and negative data construction pipeline to diversify the dataset. For the second challenge, we propose to enhance the model's focus on visual content during decoding by refining the text and visual prompt. For the text prompt, we propose a content guided refinement strategy that performs preliminary visual content analysis to generate structured information before answering the question. Additionally, we employ a visual attention refinement strategy that highlights question-relevant visual tokens to increase the model's attention to visual content that aligns with the question. Extensive experiments demonstrate that these challenges can be significantly mitigated with our proposed dataset and techniques.

cs.CV↗

arXiv2025

Derived $p$-adic heights and the leading coefficient of the Bertolini--Darmon--Prasanna $p$-adic $L$-function

Let $E/\mathbb{Q}$ be an elliptic curve and let $p$ be an odd prime of good reduction for $E$. Let $K$ be an imaginary quadratic field satisfying the classical Heegner hypothesis and in which $p$ splits. The goal of this paper is two-fold: (1) We formulate a $p$-adic BSD conjecture for the $p$-adic $L$-function $L_{\mathfrak{p}}^{\rm BDP}$ introduced by Bertolini--Darmon--Prasanna. (2) For an algebraic analogue $F_{\mathfrak{p}}^{\rm BDP}$ of $L_{\mathfrak{p}}^{\rm BDP}$, we show that the ``leading coefficient'' part of our conjecture holds, and that the ``order of vanishing'' part follows from the expected ``maximal non-degeneracy'' of an anticyclotomic $p$-adic height. In particular, when the Iwasawa--Greenberg Main Conjecture $(F_{\mathfrak{p}}^{\rm BDP})=(L_{\mathfrak{p}}^{\rm BDP})$ is known, our results determine the leading coefficient of $L_{\mathfrak{p}}^{\rm BDP}$ at $T=0$ up to a $p$-adic unit. Moreover, by adapting the approach of Burungale--Castella--Kim, we prove the main conjecture for supersingular primes $p$ under mild hypotheses. In the $p$-ordinary case, and under some additional hypotheses, similar results were obtained by Agboola--Castella, but our method is new and completely independent from theirs, and apply to all good primes.

math.NT↗

arXiv2025

An Empirical Study on Eliciting and Improving R1-like Reasoning Models

In this report, we present the third technical report on the development of slow-thinking models as part of the STILL project. As the technical pathway becomes clearer, scaling RL training has become a central technique for implementing such reasoning models. We systematically experiment with and document the effects of various factors influencing RL training, conducting experiments on both base models and fine-tuned models. Specifically, we demonstrate that our RL training approach consistently improves the Qwen2.5-32B base models, enhancing both response length and test accuracy. Furthermore, we show that even when a model like DeepSeek-R1-Distill-Qwen-1.5B has already achieved a high performance level, it can be further refined through RL training, reaching an accuracy of 39.33% on AIME 2024. Beyond RL training, we also explore the use of tool manipulation, finding that it significantly boosts the reasoning performance of large reasoning models. This approach achieves a remarkable accuracy of 86.67% with greedy search on AIME 2024, underscoring its effectiveness in enhancing model capabilities. We release our resources at the STILL project website: https://github.com/RUCAIBox/Slow_Thinking_with_LLMs.

cs.CL↗

arXiv2025

Efficient and Distributed Large-Scale Point Cloud Bundle Adjustment via Majorization-Minimization

Point cloud bundle adjustment is critical in large-scale point cloud mapping. However, it is both computationally and memory intensive, with its complexity growing cubically as the number of scan poses increases. This paper presents BALM3.0, an efficient and distributed large-scale point cloud bundle adjustment method. The proposed method employs the majorization-minimization algorithm to decouple the scan poses in the bundle adjustment process, thus performing the point cloud bundle adjustment on large-scale data with improved computational efficiency. The key difficulty of applying majorization-minimization on bundle adjustment is to identify the proper surrogate cost function. In this paper, the proposed surrogate cost function is based on the point-to-plane distance. The primary advantages of decoupling the scan poses via a majorization-minimization algorithm stem from two key aspects. First, the decoupling of scan poses reduces the optimization time complexity from cubic to linear, significantly enhancing the computational efficiency of the bundle adjustment process in large-scale environments. Second, it lays the theoretical foundation for distributed bundle adjustment. By distributing both data and computation across multiple devices, this approach helps overcome the limitations posed by large memory and computational requirements, which may be difficult for a single device to handle. The proposed method is extensively evaluated in both simulated and real-world environments. The results demonstrate that the proposed method achieves the same optimal residual with comparable accuracy while offering up to 704 times faster optimization speed and reducing memory usage to 1/8. Furthermore, this paper also presented and implemented a distributed bundle adjustment framework and successfully optimized large-scale data (21,436 poses with 70 GB point clouds) with four consumer-level laptops.

cs.RO↗

arXiv2025

Interfacial superconductivity and a Se-vacancy ordered insulating phase in the FeSe/PbOx heterostructures

The discovery of high-temperature superconductivity in FeSe/SrTiO3 has sparked significant interests in exploring new superconducting systems with engineered interfaces. Here, using molecular beam epitaxy growth, we successfully fabricate FeSe/PbOx heterostructures and discover superconductivities in three different monolayer FeSe-related interfaces. We observe superconducting gaps of 13~14 meV in the monolayer FeSe films grown on two different phases of PbOx. Moreover, we discover a new insulating Fe10Se9 phase with an ordered $\sqrt{5}\times\sqrt{5}$ Se-vacancy structure. Our first-principles calculation suggests that this new insulating phase originates from electronic correlation. Intriguingly, an additional monolayer FeSe film grown on the insulating Fe10Se9 also exhibits superconductivity with the gap size of 5 meV. Our results suggest that the work function differences between the monolayer FeSe and the substrates, which can induce band bending and charge transfer, are crucial for the interfacial superconductivity.

cond-mat.supr-con↗

arXiv2025

Any Information Is Just Worth One Single Screenshot: Unifying Search With Visualized Information Retrieval

With the popularity of multimodal techniques, it receives growing interests to acquire useful information in visual forms. In this work, we formally define an emerging IR paradigm called \textit{Visualized Information Retrieval}, or \textbf{Vis-IR}, where multimodal information, such as texts, images, tables and charts, is jointly represented by a unified visual format called \textbf{Screenshots}, for various retrieval applications. We further make three key contributions for Vis-IR. First, we create \textbf{VIRA} (Vis-IR Aggregation), a large-scale dataset comprising a vast collection of screenshots from diverse sources, carefully curated into captioned and question-answer formats. Second, we develop \textbf{UniSE} (Universal Screenshot Embeddings), a family of retrieval models that enable screenshots to query or be queried across arbitrary data modalities. Finally, we construct \textbf{MVRB} (Massive Visualized IR Benchmark), a comprehensive benchmark covering a variety of task forms and application scenarios. Through extensive evaluations on MVRB, we highlight the deficiency from existing multimodal retrievers and the substantial improvements made by UniSE. Our work will be shared with the community, laying a solid foundation for this emerging field.

cs.CL↗

arXiv2025

Does RAG Really Perform Bad For Long-Context Processing?

The efficient processing of long context poses a serious challenge for large language models (LLMs). Recently, retrieval-augmented generation (RAG) has emerged as a promising strategy for this problem, as it enables LLMs to make selective use of the long context for efficient computation. However, existing RAG approaches lag behind other long-context processing methods due to inherent limitations on inaccurate retrieval and fragmented contexts. To address these challenges, we introduce RetroLM, a novel RAG framework for long-context processing. Unlike traditional methods, RetroLM employs KV-level retrieval augmentation, where it partitions the LLM's KV cache into contiguous pages and retrieves the most crucial ones for efficient computation. This approach enhances robustness to retrieval inaccuracy, facilitates effective utilization of fragmented contexts, and saves the cost from repeated computation. Building on this framework, we further develop a specialized retriever for precise retrieval of critical pages and conduct unsupervised post-training to optimize the model's ability to leverage retrieved information. We conduct comprehensive evaluations with a variety of benchmarks, including LongBench, InfiniteBench, and RULER, where RetroLM significantly outperforms existing long-context LLMs and efficient long-context processing methods, particularly in tasks requiring intensive reasoning or extremely long-context comprehension.

cs.CL↗

arXiv2025

Reinforced Information Retrieval

While retrieval techniques are widely used in practice, they still face significant challenges in cross-domain scenarios. Recently, generation-augmented methods have emerged as a promising solution to this problem. These methods enhance raw queries by incorporating additional information from an LLM-based generator, facilitating more direct retrieval of relevant documents. However, existing methods struggle with highly specialized situations that require extensive domain expertise. To address this problem, we present \textbf{Reinforced-IR}, a novel approach that jointly adapts a pre-trained retriever and generator for precise cross-domain retrieval. A key innovation of Reinforced-IR is its \textbf{Self-Boosting} framework, which enables retriever and generator to learn from each other's feedback. Specifically, the generator is reinforced to generate query augmentations that enhance the retriever's performance, while the retriever is trained to better discriminate the relevant documents identified by the generator. This iterative process allows the end-to-end retrieval performance to be progressively optimized using an unlabeled corpus from the target domain. In our experiment, Reinforced-IR outperforms existing domain adaptation methods by a large margin, leading to substantial improvements in retrieval quality across a wide range of application scenarios.

cs.CL↗