Searcharxiv⌕ Search

arXiv subjects

Lujia Wang

Publications and source records attributed to Lujia Wang.

At least 19 recordsLinked to original sources

Colouring signed analogues of Kneser, Schrijver, and Borsuk graphs

The Kneser signed graph $\KS(n,k)$, $k\leq n$, is the graph whose vertices are signed $k$-subsets of $[n]$ (i.e. $k$-subsets $S$ of $\{ \pm 1, \pm 2, \ldots, \pm n\}$ such that $S\cap (-S)=\emptyset$). Two vertices $A$ and $B$ are adjacent with a positive edge if $A\cap (-B)=\emptyset$ and with a negative edge if $A\cap B=\emptyset$. We prove that the balanced chromatic number of $\KS(n,k)$ is $n-k+1$. We then introduce the signed analogue of Schrijver graphs and show that choosing one representative from each antipodal pair of alternating signed $k$-sets gives a vertex-critical subgraph of $\KS(n,k)$ with respect to balanced colouring. For the Borsuk signed graph on $S^d$, we obtain balanced chromatic number $d+1$ for sufficiently small distance parameter and relate this equality to antipodal covers of the sphere. On $S^2$, spherical and projective covering bounds give three, four and six as the first successive values of the balanced chromatic number.

math.CO↗

The frustration index of fully frustrated hypercubes

The frustration index of a signed graph is the minimum number of edges meeting every negative cycle. Let $F_n$ be the frustration index of the signed $n$-cube $Q_n$ in which every square is negative. Using signings constructed from weighted cubes, we prove that for every positive integer $r$, \[ F_{r^2+1}=(r^2-r+1)2^{r^2-1}. \] When $4\mid r$, our construction also gives the exact value $F_n=(n-r-p/4)2^{n-2}$ in the six dimensions $n=r^2+1+pr/2$, $p\in\{\pm1,\pm2,\pm3\}$. In dimension nine an argument with integral eigenvectors gives $F_9=772$. By varying one weight in the construction for $r^2+1$, we obtain $F_n=(n-\sqrt n+O(n^{-1/2}))2^{n-2}$ for arbitrary $n$. The values at $n=r^2+1$ show that the error term has the best possible order. Taking the positive edges of an optimal signing gives a square-free subgraph of $Q_n$ with at least $(n+\sqrt n-O(n^{-1/2}))2^{n-2}$ edges whose complement is also square-free.

math.CO↗

A planar signed graph with circular chromatic number five

A circular $r$-coloring of a signed graph maps its vertices to a circle of circumference $r\ge2$, with the endpoints of negative edges at circular distance at least $1$ and those of positive edges at most $r/2-1$; the circular chromatic number is the infimum of such $r$. We construct a simple planar signed triangulation on $85$ vertices with circular chromatic number exactly $5$. This improves the lower bound on the supremum of the circular chromatic numbers of simple planar signed graphs from $14/3$ to $5$, giving a partial answer to a question posed by Naserasr, Wang, and Zhu.

math.CO↗

Rethinking Underwater Image Enhancement: Learning through Text-Guided Degradation Curriculum

Existing paired clean-degraded datasets have advanced learning-based underwater image enhancement, but they remain either limited in scale or difficult to acquire, and still fail to adequately cover the diverse degradations encountered in real-world underwater environments. Such scarce and insufficient data ultimately restrict the generalization of underwater image enhancement models, highlighting the need for realistic and controllable degradation modeling to better capture real underwater conditions and improve enhancement robustness. To this end, we propose a unified degradation-aware training framework for the underwater image enhancement task. A Text-Conditioned Degradation Augmentation Pipeline uses a text-guided diffusion model to synthesize diverse and semantically consistent degraded samples, providing paired data to support the subsequent curriculum learning strategy. The proposed Progressive Degradation Curriculum organizes training across multiple attributes and difficulty levels to promote balanced learning, and a Transmission-Guided One-Step Diffusion Model leverages this progressive process to better address the wide range of degradation severity and restoration ambiguity in underwater images. Extensive experiments on multiple underwater benchmarks demonstrate that our proposed method achieves superior performance. Our code and models will be released upon the publication of this work.

cs.CV↗

SEDualVLN: A Spatially-Enhanced Dual-System Framework for Vision-Language Navigation

Vision-Language Navigation (VLN) approaches have currently followed two primary paradigms: the end-to-end Vision-Language Model (VLM) policy fine-tuned on navigation trajectories to directly predict actions, and the zero-shot modular pipeline integrating pre-trained Multimodal Large Language Model (MLLM) for training-free generalization to unseen environments. However, end-to-end methods struggle with long-horizon navigation and lack dynamic reasoning ability, whereas zero-shot methods are constrained by limited spatial grounding for reliable planning and require substantial reasoning time. To bridge this gap, we introduce SEDualVLN, a spatially-enhanced dual-system VLN framework. System 1 is a VLM model enhanced with both global and local spatial awareness, tailored for high-frequency action generation. System 2 integrates a general MLLM with a mapping module, wherein the MLLM performs low-frequency waypoint planning by leveraging top-down views of the real-time 3D map alongside streams of rendered path images. Both systems exploit complementary spatial representations to improve the agent's spatial understanding in VLN tasks. Ultimately, they cooperate to complete the navigation task through a fast-slow coordinated approach. SEDualVLN achieves state-of-the-art performance on VLN-CE benchmarks, and further ablation studies demonstrate the effectiveness of each system and module.

cs.RO↗

LERR-Nav: Learning to Coordinate Exploration, Recovery, and Reminiscing for Object Navigation

Object Navigation in unseen environments remains a fundamental challenge in embodied AI, as it requires agents to generalize visual and spatial reasoning to novel layouts. While recent advances in large language model (LLM) and vision-language model (VLM) have enhanced semantic mapping, spatial memory, and waypoint selecting for navigation, successful navigation requires more than selecting the next promising waypoint. A robot must also recognize stalled progress, escape unproductive behaviors, and revisit overlooked target cues. Addressing these challenges requires both specialized navigation strategies and a high-level policy that coordinates their execution. Motivated by the analysis of failed navigation episodes, we propose LERR-Nav, an object navigation framework that coordinates exploration, recovery, and reminiscing through a learned high-level policy. The framework comprises four complementary navigation strategies: value-function-guided exploration, LLM-guided exploration, recovery from stalled or repetitive behavior, and memory-guided reminiscing. To support history-aware coordination, we maintain the execution memory that preserves textual summaries of all strategy executions alongside a sliding window of map sequences. To learn the coordination policy, we initialize a high-level vision-language agent from Qwen2.5-VL-7B-Instruct and train it through a two-stage curriculum. The first stage uses map-centric supervised fine-tuning (SFT) to establish spatial grounding. The second stage performs SFT on variable-length navigation segments to learn which navigation strategy to invoke next based on current maps and the hybrid execution memory. Experiments on HM3D and MP3D demonstrate state-of-the-art navigation performance, while ablation studies validate the contributions of individual components.

cs.RO↗

Have You Seen It Clearly? Uncertainty-Guided Belief Refinement for Zero-Shot Object Navigation

Zero-Shot Object Navigation (ZSON) requires an agent to locate target objects in unfamiliar environments without additional training for the test categories or environments. Recent approaches leverage large language models and vision-language models to construct spatial representations and guide exploration. However, whether the available evidence sufficiently supports a navigation choice, and how new evidence changes the search value of previously visited locations, remain insufficiently addressed. We propose RefBeliefMap, an uncertainty-guided mapping framework that distinguishes target-location belief from spatial and semantic evidence uncertainty. By jointly modeling intra-observation uncertainty and inter-observation disagreement over a sliding memory window, RefBeliefMap assesses the reliability of both evidence channels. When competing frontiers remain insufficiently resolved, these uncertainty estimates guide active observation and semantic reasoning to refine the evidence supporting target-location beliefs before navigation commitment. Building on the refined map, we introduce RefBeliefNav, which jointly evaluates unexplored frontiers and historical waypoints according to their target-finding potential, expected information gain, and navigation cost. This unified objective allows historical locations to be reconsidered as new evidence changes their search value, coordinating exploration and revisitation. Experiments on HM3D and MP3D demonstrate state-of-the-art performance among the evaluated zero-shot baselines, with ablation studies examining the contributions of modules. Real-world experiments on a Unitree Go2 quadruped robot further demonstrate the sim-to-real transferability of our method.

cs.RO↗

Disjoint unions and fixed-parameter algorithms for chromatic villainy

Chromatic villainy is the maximum, over rearrangements of optimal graph colorings, of the minimum Hamming distance to a proper coloring with the same color multiplicities. Weak villainy allows the repair to change these multiplicities. We give exact composition laws for disjoint unions and determine the eventual growth under repetition. Write $mG$ for the disjoint union of $m$ copies of $G$. For every graph $G$, with $k=χ(G)$, weak villainy satisfies $B_{\mathrm w}(mG)=mF_k(G)$ whenever $k\mid m$, where $F_k(G)$ is the maximum distance to a proper $k$-coloring over all assignments of $k$ color labels to vertices. Its bounded defect from $mF_k(G)$ is eventually periodic; we give an explicit stabilization bound and identify the least period, which divides $k$. For ordinary villainy, we obtain a computable rational slope $Λ(G)\ge F_k(G)$ such that $B(mG)=mΛ(G)+O_G(1)$, where the error is bounded independently of $m$ and eventually periodic, with a computable period and threshold. Consequently the difference between the two values is bounded exactly when $Λ(G)=F_k(G)$, and otherwise grows linearly. The slope is determined by a finite optimization using at most $k$ source assignments. We also answer the unrestricted superadditivity questions negatively: for every $C>0$, a connected graph can lose more than $C$ in weak villainy after union with a triangle, and some connected $3$-chromatic graph $G$ satisfies $B(G\sqcup G)<2B(G)-C$. Finally, both exact values on an $n$-vertex graph can be computed in time $f(B_{\mathrm w}(G))(n+1)^{O(1)}$, without either value being supplied. Threshold recognition is fixed-parameter tractable as well.

math.CO↗

Exact and extremal values of chromatic villainy

Chromatic villainy measures the largest number of vertices that must change color to repair a rearrangement of an optimal coloring while preserving every color multiplicity. Weak villainy allows different multiplicities in the repair. On the odd cycle $C_{2r+1}$, we determine the largest repair distance from arbitrary color assignments with a prescribed color-count vector (profile) $\mathbf m$ to proper colorings with any realizable target profile $\mathbf p$: it equals $2r+1-\sum_X\min\{p_X,\lceil m_X/2\rceil\}$. In particular, every feasible profile $(a,b,c)$ has ordinary and weak repair radius $\lfloor a/2\rfloor+\lfloor b/2\rfloor+\lfloor c/2\rfloor$, and $B(C_{2r+1})=r$ for $r\ge2$, answering a question of Clark et al. For arbitrary complete multipartite graphs, we give an arithmetic formula for ordinary villainy with at most $k^2$ candidates and an exact finite formula for weak villainy, where $k$ is the number of parts. Both values are fixed-parameter tractable in $k$ for binary-encoded part sizes. We also determine the maximum of each parameter among all $n$-vertex $k$-chromatic graphs. For $k\ge3$, the ordinary maximum is $\min\{n,4(n-k)\}$, except when $n=4(n-k)-1$ or $(n,k)\in\{(5,3),(7,3)\}$, when it is $n-1$. For $k\ge2$, the weak maximum is $n-\max\{\lceil n/k\rceil,2k-n\}-\varepsilon$, where $\varepsilon\in\{0,1\}$ is determined explicitly by $n$ and $k$. Both extrema are attained by complete multipartite graphs, and a weak extremizer needs at most three distinct part sizes.

math.CO↗

Efficient Multi-Modal Planning with Reward-Guided Preference Optimization for Autonomous Driving

Safe and efficient trajectory planning is essential in autonomous driving. However, existing end-to-end approaches often fall short in both computational efficiency and safety guarantees. Methods based on imitation learning suffer from causal confusion, while rule-based scoring approaches often incur heavy computational overhead and suffer from objective misalignment. Additionally, preference-based methods rely on strict pairwise annotations, limiting data utilization. To overcome these limitations, we propose EMPlan, an efficient multi-modal trajectory planning method powered by reward-guided fine-tuning. We design a hybrid architecture that combines sparse anchors with an offset refinement module for efficient multi-modal trajectory prediction. Sparse anchors provide coarse trajectory candidates with low latency, which are subsequently refined by the offset module for higher prediction accuracy. To enhance safety without incurring additional inference costs, we adopt a two-stage training paradigm consisting of pretraining and reward-guided fine-tuning. During fine-tuning, we leverage rule-based reward signals and unpaired preference supervision to refine the pretrained policy toward safer trajectory selection. We evaluate EMPlan on the non-reactive NAVSIM benchmark, where it strikes a favorable balance between planning accuracy and efficiency, demonstrating superior performance under real-time constraints.

cs.RO↗

MVG-WAM: Multiple View Geometry-Aware World-Action Modeling for Robotic Manipulation

World-Action Models (WAMs) couple visual dynamics with action prediction, bringing the rich priors of pretrained video models to robotic manipulation. However, their multi-view interfaces typically tile images or concatenate tokens, leaving the geometric relationships among synchronized cameras implicit. This makes it harder to connect global scene context with the local geometry required for interaction. We introduce the Multi-View Geometry-Aware World-Action Model (MVG-WAM), which organizes these observations as related projections of one physical world rather than separate images on a canvas. Our model combines an epipolar-constrained global state with view-indexed geometric states jointly inferred from synchronized observations. Camera-aware routing supplies each video region with its corresponding geometric context and the shared global state, explicitly structuring the representation used for action prediction. We further ground the geometry-aware representation in metric scale through multi-horizon future-depth supervision, without requiring depth decoding during action rollout. MVG-WAM achieves average success rates of 99.1% on LIBERO and 92.07% on RoboTwin 2.0, demonstrating competitive performance across both benchmarks. Real-world experiments on Cobot Magic further demonstrate a 91.3% success rate across 150 trials spanning three manipulation tasks.

cs.RO↗

SLIP-VLA: Single-Step Latent Imagination for Policy Learning in Vision-Language-Action Models

Vision-Language-Action models are increasingly effective for robotic manipulation, yet most predict actions directly from current observations without explicitly modeling future scene evolution. Recent methods introduce future prediction to improve action generation, but dense future modeling often requires expensive iterative denoising, while one-step alternatives can underperform their multi-step counterparts. To reconcile efficient future modeling with strong action performance, we present SLIP-VLA, a policy learning framework that equips VLA models with a Single-Step Latent Imagination for future-aware action prediction. SLIP-VLA obtains temporally dense future latent representations with a single denoising update, and we improve the perceptual sufficiency of these representations by aligning intermediate latents with future geometric and semantic features. We further improve their control sufficiency through action-conditioned latent world modeling and inverse dynamics modeling, explicitly coupling latent transitions with robot actions. SLIP-VLA achieves state-of-the-art performance across diverse simulation benchmarks and real-world manipulation tasks, while its single-step latent imagination takes only 12 ms.

cs.RO↗

TactileReflex: Noise-Statistics-Driven Vision-Tactile Reflex Control for Force-Sensitive Manipulation

Manipulating fragile deformable containers, such as disposable plastic cups filled with liquid, demands real-time grip-force adaptation within an extremely narrow force margin: insufficient force causes slip, while excessive force irreversibly deforms the thin wall. Existing approaches struggle to achieve such force-sensitive manipulation tasks. We propose a noise-statistics-based calibration-driven reflex control paradigm with vision-based tactile sensing: by analyzing the sensor's intrinsic noise characteristics (via a brief static-hold-and-unload protocol), we directly derive all controller thresholds, eliminating external force calibration, trial-and-error manual tuning, or material-specific physical models. Instantiating this paradigm, we present TactileReflex, a three-channel closed-loop controller that extracts three image-level proxies, shear intensity ($S_y$), contact intensity ($F_n$), and center of pressure ($C$), from dual visuo-tactile sensors and drives prioritized reflex channels at ~12 Hz for slip suppression, weight-adaptive release, and force protection. Each channel closes the loop directly on its proxy via noise-derived thresholds. Ablation demonstrates that only the full three-channel system is able to prevent irreversible container deformation (5/5 success vs. at most 1/5 for partial configurations). In a dynamic pouring task, fixed-effort baselines fail in all 10 attempts due to pose drift, while TactileReflex achieves 9/10 success across two water volumes. As a self-contained and interpretable controller, TactileReflex can serve as a plug-and-play safety layer beneath high-level manipulation pipelines, including haptic-free VR teleoperation and vision-language-action (VLA) policies.

cs.RO↗

Obstructions to coloring arithmetic graphs

The arithmetic graph $B_n$ joins distinct $a,b\in\N$ when $\max(a,b)/\gcd(a,b)\le n$. We prove $χ(B_{205})=206$, disproving the conjecture that $χ(B_n)=n$ for every $n$, equivalently the Rainbow Cascades Conjecture. The proof reduces an arbitrary tiling by the arithmetic exponent tile to a periodic tiling, then to two families of finite quotients, which are excluded using exact computations. We also construct a $208$-coloring using $\Z_{104}\times\Z_2$ and prove $212\leχ(B_{211})\le213$. The lower bound at $211$ follows from prime-cardinality tiling rigidity and the published nonexistence of a cyclic logarithm of length $211$; we give a direct proof of the required rigidity statement. Finally, we record the equivalence with the List Cascade Coloring Conjecture and the conjecture on ironic decorations, and deduce finite graph counterexamples to both. The least $n$ with $χ(B_n)>n$ is either $195$ or $205$; determining which remains open.

math.CO↗

Balanced even cycles in signed graphs:Turán bounds, double covers, and parity obstructions

We study Turán problems for balanced even cycles in simple signed graphs, where signed subgraphs are considered up to switching. For every balanced bipartite signed graph, the signed and ordinary Turán numbers differ by at most a factor of two. Our main structural results concern the underlying graphs that admit a signing in which every $2k$-cycle is unbalanced. We characterize these graphs by the absence of an odd dependence among their $2k$-cycle incidence vectors, give a cohomological formulation, and construct subgraph-minimal obstructions of arbitrarily large order. In particular, there is no finite forbidden-subgraph characterization. We also give an exact closed-walk criterion for cycles in double covers and derive a direct signed breadth-first-search upper bound. As applications, we prove \[ \hex(n,C_{+4})=\left(\frac{\sqrt2}{2}+o(1)\right)n^{3/2} \] and study the signed hexagon number $R_6(n)=\hex(n,\{C_{-3},C_{+6}\})$. We characterize the underlying graphs counted by $R_6$ and express it as an extremal problem for ordinary $C_6$-free graphs with a prescribed involution. For every sufficiently large $n$, we construct examples with $Ω(n^{4/3})$ edges, and we give an equivariant construction attaining the coefficient obtained from the Füredi--Naor--Verstraëte lower bound by double-cover transfer. Finally, we give $n$-vertex $C_{+10}$-free signed graphs with $Ω(n^{6/5})$ edges and use octagon examples to illustrate the limitations of theta-freeness as a signing criterion.

math.CO↗

DreamTrajectory: Trajectory-Guided Action Generation with World Model Alignment for Mobile Manipulation

Mobile manipulation requires a robot to coordinate base and arm motion under continuously changing viewpoints and contact conditions, within an action space far larger than that of fixed-base manipulation. Existing Vision-Language-Action (VLA) policies are limited in two respects. (i)They map observations directly to whole-body action chunks, searching this large action space without an explicit task-space motion plan, which makes coordinated base--arm prediction imprecise. (ii)They execute the predicted chunk open-loop, without checking whether the actions can realize the motion the policy intended, so control errors and unmodeled contacts accumulate into a gap between planned and realized motion. We present DreamTrajectory, a trajectory-guided framework for language-conditioned mobile manipulation that introduces one component for each limitation. Addressing(i), DreamTrajectory jointly predicts an intention-level end-effector trajectory and a whole-body action chunk in a single action expert, so that the trajectory explicitly guides base--arm action generation instead of remaining implicit. Addressing(ii), a lightweight trajectory world model predicts the trajectory that a candidate action chunk would induce, and a test-time search--predict--score procedure selects the candidate best aligned with the planned trajectory. On MS-HAB, trajectory guidance raises average success from 32.3% to 47.5% and test-time refinement further to 54.8%, with the largest gains on contact-rich articulated-object tasks. On three real-world mobile manipulation tasks, the corresponding average success rates are 63.3%, 81.7%, and 90.0%.

cs.RO↗

LIRA: Local Cross-Layer Information Routing for Vision-Language-Action Decoding

Vision-Language-Action (VLA) models transform representations from pretrained vision-language models (VLMs) into robot actions, yet the interface that routes intermediate VLM features into action decoders remains underexplored. Existing designs either expose only a narrow part of the representation hierarchy or rigidly match each decoder block to one VLM layer, restricting access to complementary task evidence across depths. We introduce LIRA, a local cross-layer action-conditioning mechanism that formulates VLM-to-action conditioning as depth-aware information routing. LIRA operates on task-token features and LIRA Query features derived from intermediate VLM states, then assigns each Parallel Fusion Block a depth-aligned local window centered on its corresponding VLM layer. Parallel Fusion Blocks aggregate neighboring LIRA Query features and integrate them with task-token features and proprioceptive inputs before action prediction. This routing interface leaves the backbone architecture, action decoder, and supervised training recipe unchanged. Across LIBERO, LIBERO-Plus, CALVIN ABC$\rightarrow$D, and real-world manipulation, LIRA improves the principal aggregate metrics over the VLA-Adapter baseline under the same 0.5B-parameter configuration. In zero-shot transfer to LIBERO-Plus, LIRA increases average success from 59.1% to 78.0%, an 18.9-point gain indicating improved robustness under controlled distribution shifts.

cs.RO↗

Winding number and circular coloring

In 1996, Youngs proved a surprising theorem that quadrangulations of the projective plane could never have chromatic number exactly 3. This sparked a lot of interest, and the result has been further developed in many directions over the past decades. For example, the result is strengthened by considering the circular chromatic number, which is a real-valued lower bound on the chromatic number. The circular chromatic number of a quadrangulation cannot be in the interval (2,4). This parameter allows a generalization to larger even faces, for which a similar gap exists. In this work, we place these results into a framework based on the notion of winding number using extensions of colorings to continuous mappings. This yields unified and simplified proofs of gaps in the circular chromatic number for graphs with a distinguished set of directed even cycles. This generalizes the setting of graphs embedded on surfaces where every face is even. We further establish an analogous gap phenomenon when all faces are of a given odd length, previously known only in the case of triangulations. For example, we conclude that if G is a graph embedded on the projective plane such that all faces are 5-cycles, then either its circular chromatic number is 5/2 or at least 3, the former being the case only if G is Eulerian and every noncontractible facial walk is of odd length...

math.CO↗