arXiv subjects

Seunghun Lee

Publications and source records attributed to Seunghun Lee.

At least 19 recordsLinked to original sources

arXiv2026

PLED-VINS: A Point-Line Event-Based Visual Inertial SLAM for Dynamic Environments

Dynamic environments remain a fundamental challenge for visual SLAM, where unreliable observations from moving objects and rapid motion degrade state estimation accuracy. Although event cameras preserve fine-grained spatio-temporal information, most existing event-based SLAM frameworks still assume static scenes and lack approaches to estimate the reliability of features. To this end, we propose PLED-VINS, a monocular event camera-based visual-inertial SLAM framework that enables robust state estimation in dynamic environments. We propose an entropy-recency score map to characterize the temporal reliability of both point and line features based on event temporal statistics. Concurrently, geometric reliability is estimated via a unified point-line robust bundle adjustment. Building upon these, we design an adaptive weighting strategy that fuses temporal and geometric reliability, including motion-conditioned reliability modeling for line features, to suppress unreliable observations. Experimental results demonstrate that PLED-VINS improves state estimation on the evaluated dynamic sequences with moving objects.

cs.RO↗

arXiv2026

Tensor-Network Algorithm for Many-Body Trace Norms

Trace norms are fundamental to quantum information theory, yet in many-body systems their evaluation remains a major computational bottleneck, as it generally requires diagonalizing exponentially large operators. Here, we overcome this bottleneck by introducing a controlled tensor-network algorithm for estimating the trace norm of matrix product operators without full diagonalization. The key idea is to combine Zolotarev's rational approximation to the sign function with a variational formulation solved using a density-matrix-renormalization-group-like algorithm. The resulting approximation is systematically improvable, with its accuracy controlled by the rational approximation parameters and the spectral weight near zero. Beyond the reach of exact diagonalization, we demonstrate controlled trace-norm calculations for entanglement negativity, quantum fidelity and quantum Fisher information, achieving substantially improved accuracy over polynomial-based Lanczos approaches. Our results establish trace-norm-based quantities as practical tensor-network observables, opening a route toward tensor-network studies of quantum information in mixed states.

quant-ph↗

arXiv2026

Robust Mixed-State Cluster States and Spurious Topological Entanglement Negativity

We investigate 1D and 2D cluster states under local decoherence to assess the robustness of their mixed-state subsystem symmetry-protected topological (SSPT) order. By exactly computing fidelity correlators via dimensional reduction of effective statistical mechanics models, we pinpoint the critical error rate for strong-to-weak spontaneous breaking of strong subsystem symmetry. Without resorting to the replica trick, we demonstrate that mixed-state SSPT order remains remarkably robust up to the maximal decoherence rate when noise respects strong subsystem symmetry. Furthermore, we propose that the mixed-state SSPT order can be detected by a constant correction to the area-law scaling of entanglement negativity, termed spurious topological entanglement negativity. This also highlights that topological entanglement negativity, a widely used diagnostic for mixed-state topological order, is generally not invariant under finite-depth quantum channels.

quant-ph↗

arXiv2026

Mixed-State Topological Order under Coherent Noise

Mixed-state phases of matter under local decoherence have recently garnered significant attention due to the ubiquitous presence of noise in current quantum processors. One of the key issues is understanding how topological quantum memory is affected by realistic coherent noise, such as random rotation noise and amplitude-damping noise. In this work, we investigate the intrinsic error threshold of the two-dimensional toric code (TC), a paradigmatic topological quantum memory, under these types of coherent noise by employing both analytical and numerical methods based on the doubled-Hilbert-space formalism. A connection between the mixed-state phase of the decohered TC and a non-Hermitian Ashkin-Teller-type statistical-mechanics model is established, and the mixed-state phase diagrams under the coherent noise are obtained. We find remarkable stability of mixed-state topological order under random rotation noise with axes near the $Y$-axis of qubits. We also identify intriguing extended critical regions at the phase boundaries, highlighting a connection with non-Hermitian physics. We argue that these phase boundaries provide upper bounds for the intrinsic error threshold, beyond which quantum error correction becomes impossible. We complement these findings by estimating the error thresholds for random rotation noise under standard quantum error correction, thereby providing lower bounds on the intrinsic error threshold.

quant-ph↗

arXiv2026

Learning to Solve Generative ODEs Beyond the Linear Span

Diffusion and flow generative models sample by integrating a learned ODE, but high quality still requires many sequential model evaluations. Solver learning reduces this cost by adapting scalar coefficients, timesteps, or both, while keeping the backbone model fixed. In this work, we identify a structural bottleneck in this update family: each step remains span-limited. Since the scalar-coefficient update lies in the span of buffered velocity evaluations, it can fit only the in-span component while leaving any out-of-span residual unreachable by scalar recombination alone. We propose SpanLift, a lightweight neural solver that augments scalar-coefficient updates with a spatial residual operator. SpanLift keeps a fixed base solver as an in-span prior and learns a spatial residual operator over the state and velocity buffer. The operator is trained by endpoint teacher matching, preserves the pretrained backbone, and adds no model NFEs. Empirically, the learned correction transfers across base solvers and is predominantly out-of-span. Across pixel-space diffusion, latent flow matching, and precipitation nowcasting, SpanLift achieves state-of-the-art few-step sampling. With only 3 NFE, it improves CIFAR-10 FID from 8.16 to 5.69 and ImageNet FID from 17.37 to 11.83.

cs.CV↗

arXiv2026

Low-Latency Edge LLM Handover via Joint KV Cache Transfer and Token Prefill

Edge deployment of large language models (LLMs) can reduce latency for interactive services, but mobility introduces service interruptions when an user equipment (UE) hands over between base stations (BSs). To promptly resume decoding, the target-side edge server must recover the UE context state, which can be provisioned either by token forwarding followed by prefill computation or by direct key-value (KV) cache transmission over backhaul. This paper proposes a unified handover (HO) design that jointly selects the prefill length and schedules backhaul KV cache delivery to minimize the worst-user LLM HO delay for multiple UEs. The resulting scheme admits a tractable step-wise solution with explicit feasibility conditions and a constructive rate-scheduling policy. Simulations show that the proposed method consistently outperforms baselines across a wide range of backhaul capacities, prefill speeds, and context sizes, providing practical guidelines for mobility-aware Edge LLM token streaming.

eess.SP↗

arXiv2026

CVA: Context-aware Video-text Alignment for Video Temporal Grounding

We propose Context-aware Video-text Alignment (CVA), a novel framework to address a significant challenge in video temporal grounding: achieving temporally sensitive video-text alignment that remains robust to irrelevant background context. Our framework is built on three key components. First, we propose Query-aware Context Diversification (QCD), a new data augmentation strategy that ensures only semantically unrelated content is mixed in. It builds a video-text similarity-based pool of replacement clips to simulate diverse contexts while preventing the ``false negative" caused by query-agnostic mixing. Second, we introduce the Context-invariant Boundary Discrimination (CBD) loss, a contrastive loss that enforces semantic consistency at challenging temporal boundaries, making their representations robust to contextual shifts and hard negatives. Third, we introduce the Context-enhanced Transformer Encoder (CTE), a hierarchical architecture that combines windowed self-attention and bidirectional cross-attention with learnable queries to capture multi-scale temporal context. Through the synergy of these data-centric and architectural enhancements, CVA achieves state-of-the-art performance on major VTG benchmarks, including QVHighlights and Charades-STA. Notably, our method achieves a significant improvement of approximately 5 points in Recall@1 (R1) scores over state-of-the-art methods, highlighting its effectiveness in mitigating false negatives.

cs.LG↗

arXiv2026

Polytopes with large transversal ratio

The transversal ratio of a polytope $P$ is the minimum proportion of vertices of $P$ required to intersect each facet of $P$. The weak chromatic number of $P$ is the minimum number of colors required to color the vertices of $P$ so that no facet is monochromatic. We will construct an infinite family of $d$-polytopes for each $d\geq 5$ whose transversal ratio approaches 1 as the number of vertices grows. In particular, this implies that the weak chromatic number for $d$-polytopes is unbounded for each $d\geq 5$. The previous best known lower bounds on the supremum of the transversal ratio for $d$-polytopes for $d\geq 5$ were 2/5 for odd $d$ by Novik and Zheng, and 1/2 for even $d$ by Holmsen, Pach, and Tverberg. In the case of simplicial $(d-1)$-spheres, the best known lower bounds were 1/2 for $d=5$ and $6/11$ for $d=6$ by Novik and Zheng.

math.CO↗

arXiv2026

On colorings of hypergraphs embeddable in $\mathbb{R}^d$

The (weak) chromatic number of a hypergraph $H$, denoted by $χ(H)$, is the smallest number of colors required to color the vertices of $H$ so that no hyperedge of $H$ is monochromatic. For every $2\le k\le d+1$, denote by $χ_L(k,d)$ (resp. $χ_{PL}(k,d)$) the supremum $\sup_H χ(H)$ where $H$ runs over all finite $k$-uniform hypergraphs such that $H$ forms the collection of maximal faces of a simplicial complex that is linearly (resp. PL) embeddable in $\mathbb{R}^d$. Following the program by Heise, Panagiotou, Pikhurko and Taraz, we improve their results as follows: For $d \geq 3$, we show that A. $χ_L(k,d)=\infty$ for all $2\le k\le d$, B. $χ_{PL}(d+1,d)=\infty$ and C. $χ_L(d+1,d)\ge 3$ for all odd $d\ge 3$. As an application, we extend the results by Lutz and Møller on the weak chromatic number of the $s$-dimensional faces in the triangulations of a fixed triangulable $d$-manifold $M$: D. $χ_s(M)=\infty$ for $1\leq s \leq d$.

math.CO↗

arXiv2026

Bridging Geometric and Semantic Foundation Models for Generalized Monocular Depth Estimation

We present Bridging Geometric and Semantic (BriGeS), an effective method that fuses geometric and semantic information within foundation models to enhance Monocular Depth Estimation (MDE). Central to BriGeS is the Bridging Gate, which integrates the complementary strengths of depth and segmentation foundation models. This integration is further refined by our Attention Temperature Scaling technique. It finely adjusts the focus of the attention mechanisms to prevent over-concentration on specific features, thus ensuring balanced performance across diverse inputs. BriGeS capitalizes on pre-trained foundation models and adopts a strategy that focuses on training only the Bridging Gate. This method significantly reduces resource demands and training time while maintaining the model's ability to generalize effectively. Extensive experiments across multiple challenging datasets demonstrate that BriGeS outperforms state-of-the-art methods in MDE for complex scenes, effectively handling intricate structures and overlapping objects.

cs.CV↗

arXiv2025

On an extension problem on the moment curve

We show that for $2\le d\le 4$, every finite geometric simplicial complex $Δ$ in $\mathbb{R}^d$ with vertices on the moment curve can be extended to a triangulation $T$ of the cyclic polytope $C$ where $Δ, T$ and $C$ all have the same vertex set. Further, for $d\ge 5$ we construct for every $n\ge d+3$ complexes $Δ$ on $n$ vertices for which no such triangulations $T$ exist. Our result for $d=4$ has the following novel algebraic application, due to a correspondence by Oppermann and Thomas (JEMS, 2012): every maximal rigid object in $\mathcal{O}_{A_n^{2}}$ is cluster tilting, where $\mathcal{O}_{A_n^δ}$ denotes a higher dimensional cluster category introduced by Oppermann and Thomas for $A_n^δ$, where $A_n^δ$ denotes a higher Auslander algebra of linearly oriented type $A$.

math.CO↗

arXiv2025

PRESTO: Preimage-Informed Instruction Optimization for Prompting Black-Box LLMs

Large language models (LLMs) have achieved remarkable success across diverse domains, due to their strong instruction-following capabilities. This has led to increasing interest in optimizing instructions for black-box LLMs, whose internal parameters are inaccessible but widely used due to their strong performance. To optimize instructions for black-box LLMs, recent methods employ white-box LLMs to generate candidate instructions from optimized soft prompts. However, white-box LLMs often map different soft prompts to the same instruction, leading to redundant queries. While previous studies regarded this many-to-one mapping as a structure that hinders optimization efficiency, we reinterpret it as a useful prior knowledge that can accelerate the optimization. To this end, we introduce PREimage-informed inSTruction Optimization (PRESTO), a novel framework that leverages the preimage structure of soft prompts for efficient optimization. PRESTO consists of three key components: (1) score sharing, which shares the evaluation score with all soft prompts in a preimage; (2) preimage-based initialization, which selects initial data points that maximize search space coverage using preimage information; and (3) score consistency regularization, which enforces prediction consistency within each preimage. By leveraging preimages, PRESTO achieves the effect of effectively obtaining 14 times more scored data under the same query budget, resulting in more efficient optimization. Experimental results on 33 instruction optimization tasks demonstrate the superior performance of PRESTO. Code is available at https://github.com/mlvlab/PRESTO

cs.LG↗

arXiv2025

On conflict-free colorings of cyclic polytopes and the girth conjecture for graphs

We study the conflict-free chromatic number of hypergraphs derived from the family of facets of $d$-dimensional cyclic polytopes with $n$ vertices. While in odd dimensions $d$ the problem is easy, for even dimensions the problem becomes very difficult and exhibits interesting connections to extremal graph theory. We provide sharp asymptotic bounds for the conflict-free chromatic number in several small even dimensions and non-trivial upper and lower bounds for general even dimensions. The main purpose of this paper is revealing a surprising relation between conflict-free colorings and the celebrated Erdős girth conjecture, opening new avenues for future research.

math.CO↗

arXiv2025

Temporal Grounding as a Learning Signal for Referring Video Object Segmentation

Referring Video Object Segmentation (RVOS) aims to segment and track objects in videos based on natural language expressions, requiring precise alignment between visual content and textual queries. However, existing methods often suffer from semantic misalignment, largely due to indiscriminate frame sampling and supervision of all visible objects during training -- regardless of their actual relevance to the expression. We identify the core problem as the absence of an explicit temporal learning signal in conventional training paradigms. To address this, we introduce MeViS-M, a dataset built upon the challenging MeViS benchmark, where we manually annotate temporal spans when each object is referred to by the expression. These annotations provide a direct, semantically grounded supervision signal that was previously missing. To leverage this signal, we propose Temporally Grounded Learning (TGL), a novel learning framework that directly incorporates temporal grounding into the training process. Within this frame- work, we introduce two key strategies. First, Moment-guided Dual-path Propagation (MDP) improves both grounding and tracking by decoupling language-guided segmentation for relevant moments from language-agnostic propagation for others. Second, Object-level Selective Supervision (OSS) supervises only the objects temporally aligned with the expression in each training clip, thereby reducing semantic noise and reinforcing language-conditioned learning. Extensive experiments demonstrate that our TGL framework effectively leverages temporal signal to establish a new state-of-the-art on the challenging MeViS benchmark. We will make our code and the MeViS-M dataset publicly available.

cs.CV↗

arXiv2025

Colorful circuits and colorful topes in oriented matroids

We provide a short proof of a conic version of the colorful Carathéodory theorem for oriented matroids. Holmsen's extension of the colorful Carathéodory theorem to oriented matroids (Advances in Mathematics, 2016) already encompasses several generalizations of the original result, but not its conic version. Our approach relies on a common generalization of Sperner's lemma and Meshulam's lemma-two closely related results from combinatorial topology that have found a number of applications in discrete geometry and combinatorics. This generalization may be of independent interest. Using a similar approach, we also establish the following colorful theorem for topes, whose special geometric case had not been considered before: Given $n$ topes from a uniform oriented matroid with $n$ elements, if they agree on some element, then there is a way to select a distinct element from each tope, together with its sign, so as to form another tope of the oriented matroid. Motivated by this theorem, we further explore other conditions leading to the same conclusion.

math.CO↗

arXiv2025

Latest Object Memory Management for Temporally Consistent Video Instance Segmentation

In this paper, we present Latest Object Memory Management (LOMM) for temporally consistent video instance segmentation that significantly improves long-term instance tracking. At the core of our method is Latest Object Memory (LOM), which robustly tracks and continuously updates the latest states of objects by explicitly modeling their presence in each frame. This enables consistent tracking and accurate identity management across frames, enhancing both performance and reliability through the VIS process. Moreover, we introduce Decoupled Object Association (DOA), a strategy that separately handles newly appearing and already existing objects. By leveraging our memory system, DOA accurately assigns object indices, improving matching accuracy and ensuring stable identity consistency, even in dynamic scenes where objects frequently appear and disappear. Extensive experiments and ablation studies demonstrate the superiority of our method over traditional approaches, setting a new benchmark in VIS. Notably, our LOMM achieves state-of-the-art AP score of 54.0 on YouTube-VIS 2022, a dataset known for its challenging long videos. Project page: https://seung-hun-lee.github.io/projects/LOMM/

cs.CV↗

arXiv2025

CAVIS: Context-Aware Video Instance Segmentation

In this paper, we introduce the Context-Aware Video Instance Segmentation (CAVIS), a novel framework designed to enhance instance association by integrating contextual information adjacent to each object. To efficiently extract and leverage this information, we propose the Context-Aware Instance Tracker (CAIT), which merges contextual data surrounding the instances with the core instance features to improve tracking accuracy. Additionally, we design the Prototypical Cross-frame Contrastive (PCC) loss, which ensures consistency in object-level features across frames, thereby significantly enhancing matching accuracy. CAVIS demonstrates superior performance over state-of-the-art methods on all benchmark datasets in video instance segmentation (VIS) and video panoptic segmentation (VPS). Notably, our method excels on the OVIS dataset, known for its particularly challenging videos. Project page: https://seung-hun-lee.github.io/projects/CAVIS/

cs.CV↗

arXiv2025

Low-Complexity Semantic Packet Aggregation for Token Communication via Lookahead Search

Tokens are fundamental processing units of generative AI (GenAI) and large language models (LLMs), and token communication (TC) is essential for enabling remote AI-generate content (AIGC) and wireless LLM applications. Unlike traditional bits, each of which is independently treated, the semantics of each token depends on its surrounding context tokens. This inter-token dependency makes TC vulnerable to outage channels, where the loss of a single token can significantly distort the original message semantics. Motivated by this, this paper focuses on optimizing token packetization to maximize the average token similarity (ATS) between the original and received token messages under outage channels. Due to inter-token dependency, this token grouping problem is combinatorial, with complexity growing exponentially with message length. To address this, we propose a novel framework of semantic packet aggregation with lookahead search (SemPA-Look), built on two core ideas. First, it introduces the residual semantic score (RSS) as a token-level surrogate for the message-level ATS, allowing robust semantic preservation even when a certain token packet is lost. Second, instead of full search, SemPA-Look applies a lookahead search-inspired algorithm that samples intra-packet token candidates without replacement (fixed depth), conditioned on inter-packet token candidates sampled with replacement (fixed width), thereby achieving linear complexity. Experiments on a remote AIGC task with the MS-COCO dataset (text captioned images) demonstrate that SemPA-Look achieves high ATS and LPIPS scores comparable to exhaustive search, while reducing computational complexity by up to 40$\times$. Compared to other linear-complexity algorithms such as the genetic algorithm (GA), SemPA-Look achieves 10$\times$ lower complexity, demonstrating its practicality for remote AIGC and other TC applications.

eess.SP↗