Papers

Recent progress in deep learning has significantly advanced facial landmark detection. However, most existing methods process features in a spatial-domain manner under a dataset-specific training paradigm, which overlooks the fact that facial landmark detection is inherently geometry-driven and sensitive to frequency variations, thereby limiting cross-dataset generalization under complex scenarios and hindering the development of a facial landmark detection model. To address this issue, we propose FreqFLD, a frequency-modulated framework towards All-in-One facial landmark detection. Specifically, FreqFLD introduces a Frequency Modulation Module (FreqMoM) to explicitly induce the frequency prior by decoupling and modulating low- and high-frequency components, which is then injected into subsequent feature modeling to enable balanced modeling of global facial structure and local landmark details. Furthermore, FreqFLD employs a Frequency-Modulated Mixture-of-Experts (FreqMoE), with expert selection adaptively conditioned on frequency-modulated priors, enabling flexible modeling of heterogeneous facial landmark patterns under diverse and challenging scenarios. To regularize frequency-consistent modeling under the All-in-One paradigm, we further introduce a Frequency-Consistent Routing (FreqCR) loss, which constrains the routing and assignment of frequency-aware experts to promote balanced expert utilization across diverse facial scenarios, thereby enabling stable expert specialization and achieving robust facial landmark detection. Extensive experiments demonstrate that the proposed FreqFLD achieves comparable performance on popular datasets. The code is available at: https://github.com/jkj1059657014/FreqFLD.

With the rapid spread of news on social media, understanding the propagation of misinformation is becoming increasingly important. One factor that affects individuals' vulnerability to false information is their ideological predisposition. Despite the large number of agent-based models that focus on social influence as a driver of the spread of false claims, they often fail to explicitly integrate personal ideological biases into belief formation. In this work, we explore how misinformation spreads through the interaction between individuals' ideological biases and social influence. Our model accounts for both the strength of individuals' ideological biases and the extent to which a false claim aligns with their ideology. Social influence modifies the effects of ideological intensity and false claim alignment through network interactions. Notably, the influence of neighbours' ideological intensity on belief is strongly affected by how well those neighbours are connected to one another. These results highlight the importance of considering both network structure and personal ideological biases when modelling misinformation propagation.

This paper introduces an evolutionary computation approach for consensus in structural Bayesian Network (BN) fusion under the constraint of limited treewidth. The consensus BN aims to reconcile multiple input BNs into a single one that retains key structural features present in the original networks. Treewidth, a graph-based parameter associated with computationally tractable inference, is utilized to restrict the complexity of the resulting network. A genetic algorithm is proposed to look for a BN that codifies as much information about the unrestricted fusion as possible while ensuring the treewidth restriction. Experimental evaluation demonstrates the genetic algorithm's ability to obtain consensus BNs with limited treewidth, providing a valuable tool for aggregating information from diverse sources while returning a computationally actionable model.

Affine Grassmann codes, introduced by Beelen, Ghorpade and Høholdt, are linear codes over $\mathbb{F}_q$ obtained by evaluating linear combinations of minors of a generic matrix. The weight spectrum of the affine Grassmann code $C^{\mathbb{A}}(2,m)$ was determined by Piñero and Singh via a case analysis on the rank of an associated alternating matrix. We give an independent and more streamlined derivation, valid for all $m\ge4$ and every prime power $q$, in which each codeword is written in a compact matrix form and its Hamming weight is expressed through a single closed formula involving two explicit affine subspaces and their intersection.

Open-loop multilegged locomotion over rough terrain has been interpreted as matter transport over a noisy channel: leg-ground interactions are discrete basic active contacts, terrain deletes or perturbs those contacts, and spatial redundancy concentrates the resulting thrust and arrival time. That construction is repetition-like because every module carries the same scalar locomotion task. It consequently provides neither a positive task rate nor a decoder that changes with the surviving contact set. Here we formulate locomotion instead as a quantized finite-frame expansion with erasures. A d-dimensional body-level command is mapped into N>d heterogeneous local contact commands. Rough terrain erases or corrupts frame coefficients, while a contact-gated compliant morphology physically realizes the weighted active-subframe decoder. For a linear-Gaussian model, mechanical equilibrium is exactly the posterior mean, tangent stiffness is posterior precision, and mechanical compliance is posterior covariance. Equal-norm Parseval frames are shown to be minimax optimal against one missing contact, two-contact robustness is governed by frame coherence, and a harmonic frame gives a directly realizable gait family. For independently surviving contacts of probability q, random Gaussian gait frames admit exact reconstruction at every analog dimension rate R<q, with a binomial reliability exponent, whereas recovery of arbitrary commands is impossible for R>q. Residual contact noise yields an asymptotic per-mode amplification 1/(q-R) and a vanishing mechanical stiffness margin at the threshold. An information-locomotion inequality and an exact incremental-redundancy rule direct the next gait component toward the softest task-relevant unresolved mode. The resulting analog frame-coding theorem establishes a finite relative redundancy and converse as part of a fundamental limit theory of legged locomotion.

While the GDPR and ePrivacy Directive mandate that consent information must be clear and accessible, most modern cookie banners remain obscured by technical jargon, vague phrasing, and frequent content overload or underload. This feasibility study investigates the impact of applying plain language (Einfache Sprache) to cookie banners within the IAB Transparency & Consent Framework (TCF). In our study, we analysed cookie banner texts from 200 websites, using AI-based mapping to categorise extracted content into standardised processing purposes. By substituting complex legal terms with simplified descriptions, we successfully demonstrated that the comprehension barrier can be lowered from a college-graduate level to a 7th-grade level. However, the effectiveness of plain language is inherently constrained by the informativeness of the original content; it cannot compensate for banners that omit legally required details. We conclude that while plain language is a vital tool for digital accessibility, it must be paired with standardised implementation guidelines to ensure that cookie banners are both readable and informative.

We study the vantage point selection problem, introduced by Ashvinkumar, Chowdhury, Gao, Goswami, Mitchell, and Polishchuk [WADS'25] to model the problem of estimating bottleneck capacities on the Internet. The input is a weighted undirected graph with unique shortest paths where every edge has a distinct unknown capacity. When the algorithm queries a vertex $v$, it reveals the minimum-capacity edge on the shortest path from $v$ to every other vertex reachable from $v$. The goal is to maximize the total number of revealed edges. The quality of an algorithm is measured by its competitive ratio against an optimal algorithm that knows all edge capacities a priori. We first consider the foundational single-query setting, where both the algorithm and the optimal algorithm are restricted to a single query. There is a trivial upper bound of $O(n)$ on the competitive ratio and the best known lower bound was $\tilde{\Omega}(\sqrt{n})$. We provide an algorithm and matching lower bound (up to polylogarithmic factors) showing that the best possible competitive ratio is $\tilde{\Theta}(n^{2/3})$. Furthermore, we extend our results to the general setting where the optimal algorithm is allowed $k$ queries and our algorithm is allowed $\alpha k$ queries for $\alpha\geq 1$. We present a randomized non-adaptive algorithm and matching lower bound (up to polylogarithmic factors) showing that the best possible expected competitive ratio for non-adaptive algorithms is the following surprisingly complex bound: $$ \tilde{\Theta}\left( \min\left\{ \frac{n}{\alpha k}, \max\left( \sqrt{\frac{n}{\alpha}}, \frac{n^{2/3}}{\alpha k^{1/3}} \right) \right\} \right). $$

Reusing key-value (KV) caches speeds up LLM inference by avoiding repeated computation on shared text. Standard prefix caching reuses a KV cache only when the LLM is the same and all preceding text is identical, but real workloads often break both conditions: RAG systems place different documents before the same one, agents with different system prompts read the same file or tool output, multi-agent workflows use specialized LLMs on shared material, and an updated model reads documents cached by its previous version. Because KV caches depend on both the preceding text and the model weights, direct reuse can reduce answer quality. Many methods repair or compress the reused cache, but each paper uses its own tasks, models, and cost measures, and existing benchmarks mainly test long-context processing or reuse of an unchanged prefix. We introduce KVShareArena, a benchmark and open evaluation framework for comparing them under the same conditions. KVShareArena has (1) reuse tests on 2,150 questions from three QA datasets, where the preceding text, the cache-writing LLM, or both change while the answering LLM and input stay fixed; (2) five dense and mixture-of-experts LLMs (4B-30B) and six LLM pairs where one version of an LLM reads caches written by another, for 33 model-dataset settings; (3) 11 repair and compression methods from six method classes; (4) four evaluation perspectives: answer quality, prefill computation, KV-cache memory, and latency; and (5) a common interface for adding new methods and an interactive leaderboard. Experiments yield two findings. First, both the quality loss from reuse and which repairs help depend on the LLM, even between two 8B models. Second, most repairs keep their quality when another LLM version wrote the cache, but a trained repair adapter loses quality in 12 of 18 pair-dataset tests. Code and data: https://github.com/xishi404/KVShare-Arena

Streaming multi-speaker ASR is a challenging task that must balance accuracy, latency, and efficiency while handling overlapping speech and maintaining coherent long-context modeling over extended conversations in an online fashion. We present a unified framework that categorizes streaming multi-speaker ASR into four architectural strategies based on how diarization and ASR are integrated. Using a shared pair of open-source streaming ASR and diarization models as a common foundation, we derive four multi-speaker ASR systems that differ in whether they employ multiple model instances, fine-tuning, or both. We evaluate these systems across multi-speaker accuracy, single-speaker accuracy degradation, memory footprint, and training complexity. Through this systematic architectural analysis, we clarify the design space for streaming multi-speaker ASR and provide practical guidance for selecting the most suitable approach under diverse deployment constraints.

Open-source large language models (LLMs) are increasingly competitive with closed-source models while offering transparency and the ability to run inference without exposing user inputs to a service provider. However, running large-scale models locally requires substantial computational resources. In practice, users may still resort to a third-party provider, giving rise to privacy and correctness concerns. Existing solutions that address these problems often impose substantial server overhead or introduce additional trust assumptions. In this paper, we present Maverick, a novel approach to private and verifiable LLM inference based on a protocol for delegating matrix-vector multiplication, a dominant operation in LLMs. At its core, Maverick provides, to our knowledge, the first information-theoretically sound verification protocol for matrix-vector multiplication delegation with transparent preprocessing, efficient (batch) verification, and virtually no server overhead. We combine this verification primitive with LPN-based pseudorandom masking to provide input privacy. We implement our matrix-vector delegation primitive and use it to build an end-to-end prototype of Maverick, which we evaluate on Qwen3-4B by measuring throughput in tokens per second. We evaluate client configurations with 1-8 threads. With one client thread and a CPU server using up to 128 threads, Maverick achieves throughput gains over local inference of up to 17x when privacy masks are generated online, 45x when they are precomputed, and 44x when only verification is required. With four client threads, the corresponding gains are 13x, 18x, and 17x. When server computation is no longer the bottleneck, client-side microbenchmarks with simulated network delay show speedups of 12x-20x, 34x-135x, and 38x-157x.

Persistent language agents need stored experience to remain available across time, while each answer requires evidence suited to a particular question. A superseded fact can mislead a current-state answer and still be essential for a historical query. We present RD-Forget, a training-free framework that separates what an agent stores from what it uses. A retained source archive preserves observations, and a query-conditioned memory view controls their influence on the current answer. A frozen language-model curator extracts relevant evidence, groups facts into semantic slots, and preserves the relations needed for multi-hop reasoning. Same-slot replacement links suppress superseded values in current-state contexts, while intent-aware retrieval makes earlier evidence eligible again. A rate-distortion formulation guides construction of the answer-time view within a memory budget. Experiments span conversational memory, knowledge updating, fact consolidation, long-context reasoning, and personalization under a shared answering pipeline. The results associate accurate answers with both query-relevant evidence construction and control over obsolete alternatives. Configurations without forgetting or query conditioning have the largest score deficits, while slot grouping, historical access, and relation preservation contribute complementary functions. Retaining history while selectively controlling its use offers a practical way to accommodate changing facts and future questions.

Multi-modal medical image segmentation leverages complementary diagnostic information, yet fusion can underperform single-modality baselines when spatially aligned inputs differ in quality. Here, "corruption" primarily denotes resolution-induced degradation rather than misalignment or complete modality absence, while synthetic noise is evaluated only as an auxiliary setting. We identify a critical optimization-inference inconsistency: degraded modalities can receive weak training updates yet substantially affect predictions, indicating active interference with fusion. We attribute this failure to resampling-induced feature corruption and optimization bias, where noisy features propagate through skip connections and encourage unreliable modality selection. We therefore propose CoReFuse-Med, a Corruption-aware Rebalanced Fusion framework that suppresses corruption during feature transmission and rebalances modality contributions during high-level fusion. Experiments on EPVS, BraTS, and WMH, including multiple Z-axis slice-retention ratios and an auxiliary noise test, demonstrate improved accuracy and robustness under modality-quality discrepancies. Our code is available at https://github.com/lrever/CoReFuse.

Eclipse attacks isolate a blockchain node by monopolizing its network connections. Existing attacks on Monero (NDSS'25), Bitcoin (USENIX'15/21, S&P'20) and Ethereum (WWW'26) implicitly assume that the adversary can establish inbound connections, thereby excluding a large and practically dominant class of nodes: unreachable nodes operating behind NATs. Such nodes are widely believed to enjoy stronger networks. We challenge this assumption and show that unreachability does NOT imply the expected resilience! We present the first eclipse attacks tailored to unreachable nodes in Monero's P2P network. Our attacks require no inbound access to the victim. Instead, they first poison the peerlist of reachable nodes, which subsequently act as propagation relays to contaminate unreachable nodes' whitelists. The adversary then exploits Monero's built-in outbound connection refresh logic to evict benign neighbors and eventually monopolize all outbound connections. We instantiate this strategy in two attacks: Nyx, which targets long-running unreachable nodes and achieves a complete and persistent eclipse through network-wide poisoning; and Moros, a stealthier attack that exploits the bootstrapping phase to rapidly eclipse newly joined unreachable nodes. We ethically evaluate both attacks. Nyx is validated via large-scale simulations on a Monero network constructed using the SEED Emulator, while Moros is demonstrated on the Monero mainnet against controlled targets. Our results show that unreachable nodes can be reliably driven into stable, long-lived eclipse states. We also propose countermeasures.

Agitated filter dryers (AFDs) are a crucial unit operation in small molecule manufacturing that enables simultaneous filtration, washing, and drying of active pharmaceutical ingredients. One of the key challenges in AFDs is associated with undesired agglomeration, where the presence of hard agglomerates results in off-spec products, equipment damage, and additional downstream processing. This article presents a novel mechanistic model that describes the formation of soft and hard agglomerates during agitated filter drying. By integrating population balance and multiphysics modeling, the model can accurately predict the evolution of the product temperature, moisture content, and particle size distribution, and hence quantify the extend and impact of undesired agglomeration across various operating conditions. Our proposed model-based framework enables the rational design and operation of AFDs for improving the product quality and process reliability.

Let $\lambda$ be the minimum cut value of an $n$-vertex undirected multigraph. For every fixed $\alpha>1$, we prove that there are $O(n^{\lceil2\alpha\rceil-1})$ cuts of size strictly below $\alpha\lambda$. The exponent is sharp. The proof combines splitting off and sampling with a bound on the size of nested families of vertex sets.

Large language models (LLMs) are increasingly deployed in globally used assistants, yet their default choices in culturally grounded everyday situations can systematically favour some cultures over others, affecting localisation, user trust, and equitable behaviour. Existing cultural benchmarks evaluate accuracy against a single "correct" answer, making it difficult to characterise an LLM's cultural preference prior when multiple culturally grounded responses are all valid; they also conflate default preferences with context-driven adaptation. We propose DiSCo, a distribution-first forced-choice evaluation framework that isolates default cultural priors and tests steerability via a four-level context gradient (C0–C3). Using DiSCo-Bench (304 items) derived from BLEnD spanning 12 cultures, we evaluate six diverse instruction-tuned LLMs. Default priors are heavily concentrated, with UK and US together absorbing approximately 35% of all selections despite representing only 2 of 12 cultures. Most critically, prompt-based steering consistently widens the selection gap between high- and low-resource cultures, and injecting explicit cultural facts produces negligible distributional disruption, confirming that cultural preference bias cannot be resolved through prompt-based personalisation alone.

We present and analyze an iterative refinement (IR) framework for improving the precision dependence of algorithms that combine a quantum linear system algorithm (QLSA) with quantum state tomography. Existing QLSAs achieve polylogarithmic dependence on the inverse error tolerance to prepare a quantum state encoding the solution, but extracting a classical description of such a state via tomography typically introduces a polynomial dependence on the target precision. To retain polylogarithmic dependence in the inverse error tolerance throughout the entire process, we develop an IR scheme that solves a sequence of related linear systems to progressively refine the solution while requiring only fixed-precision quantum subroutines, finally obtaining a classical description of a high-precision solution. We analyze the proposed framework under three input models: quantum-read/classical-write RAM (QRAM), linear combinations of unitaries (LCU), and sparse-access oracles. We evaluate the proposed scheme through numerical experiments on both quantum simulators and real quantum hardware. The results demonstrate that iterative refinement efficiently improves the precision of the solution and exhibits robustness to hardware noise.

Traditional software delivery assumes a static paradigm: code is constructed prior to execution and deployed as a fixed artifact. We present Agentic Just-In-Time Software Construction (A-JIT), a paradigm that replaces static binaries with dynamic, software systems that can perpetually evolve to meet changing demands. In A-JIT, an application is an integrated assembly comprising code, a runtime harness, and an embedded AI agent that continuously observes system usage and live execution traces. Much like a traditional JIT compiler specializes machine code to runtime execution paths, A-JIT specializes software logic, workflows, and tool interfaces to meet the specific needs of the end-user. By integrating synthesis directly into the ambient application lifecycle, A-JIT enables applications to dynamically construct missing implementations, generate new capabilities on the fly, and continuously adapt to end-user behavior. We demonstrate how this model supports trace-driven human-AI co-construction and opens a new design space for adaptive, self-evolving software.

Once complex, creating and deploying a new cryptocurrency has become trivial. Coin launchpads now allow users to generate a new coin with merely a few clicks, at a minimal cost. Launchpad popularity has grown in tandem with the rise of "meme coins," which usually do not offer any novel technological properties and are purely created for fun. The most prominent coin launchpad, pump.fun, has gained significant traction, grossing over 100 million USD in daily trading volume. The mass adoption of coin launchpads, however, also enables strategic actors to easily manipulate trading signals, unbeknownst to inexperienced traders who then buy certain coins, and enable these strategic actors to profit from rapid and unsustainable price increases ("pumps"). To identify such manipulations at scale, we conduct a large-scale study of pump.fun, collecting information on all 15 million coins launched in the last two years, and performing analysis on large, random samples of transaction data. We identify five classes of manipulation strategies: 1) wash trading, 2) creator address obfuscation, 3) coordinated sell, 4) copycat coins, and 5) social media manipulation. We find that strategic actors often bypass the platform interface and implement these strategies in a highly automated and low-latency fashion, by interacting directly with the blockchain. We further uncover the existence of "Market-Manipulation-as-a-service (MMaaS)," third-party tools that enable users to perform these manipulations without any technical expertise. We conclude by devising mitigations and proposing recommendations for traders, pump.fun, wallets or chain scanners, software development platforms, and regulators.

We present our system for the SHROOM-Visions 2026 shared task on character-level VLM hallucination detection. A small ($4$B-parameter) VLM is fine-tuned as a per-token classifier reading a two-token feature from its own hidden states, and is ensembled with a $\sim$400B zero-shot VLM judge at prediction time. Both components see off-the-shelf OCR of any visible in-image text. We use synthetic hallucination data generated by the large model as a source of ensemble diversity, and use validation to select feature layer, training data and OCR grounding. Our official entry reaches mean Cor $0.487$ / Cor-lbl $0.387$ on the hidden test set, placing $6$th/$28$ (EN), $6$th/$21$ (FR), $8$th/$21$ (IT) and $7$th/$22$ (ZH) on the task's primary Cor-lbl metric.

Embodied AI, including vision-language-action and world-action models, must operate reliably in the physical world. Yet methods that perform well in simulation can degrade substantially on real robots, especially in long-horizon deformable-object manipulation, where policies must track changing states and execute reliable multi-stage bimanual interactions. Existing real-robot benchmarks mainly focus on short-horizon rigid-object tasks and offer limited coverage of long-horizon deformable manipulation. We introduce RoboFolDeX, a physical-world benchmark built entirely from real-robot data, with garment folding as its primary task. Since real-robot data collection is costly, RoboFolDeX studies how heterogeneous physical experience can be reused efficiently. The benchmark is organized around four research axes: leveraging human intervention and recovery data collected during deployment; transferring data across tasks, including across garment categories and from rigid to deformable-object manipulation; reusing data across scenes with changes in lighting, background, and layout; and transferring data across robotic embodiments. RoboFolDeX provides 2,000+ hours of real-robot data spanning 20+ tasks and 10+ embodiments. We also establish a fair real-robot evaluation platform for externally submitted policies, with standardized tasks, held-out physical objects, controlled initializations, and a unified execution protocol. The platform is publicly accessible at https://ai.midea.com/#/fold-challenge. We hope RoboFolDeX will serve as a unified testbed for heterogeneous real-robot data reuse and reliable long-horizon deformable manipulation.

Directly self-starting implicit methods are attractive for transient analysis because they avoid auxiliary starting procedures while retaining the original first- or second-order governing equations. However, most existing formulations usually fix the last sub-step at the end of each time interval, which restricts the attainable order. This study develops a generalized $s$-sub-step implicit framework by releasing this constraint and treating all sub-step locations as design variables. The resulting methods admit a unified Runge–Kutta representation for both first- and second-order transient systems and preserve identical effective matrices over all sub-steps. Accuracy conditions are derived by simultaneously matching the numerical amplification factor and load operator, thereby accounting for both homogeneous and forced responses. For $s=1,~\cdots,~6$, two complementary families are obtained: $s$th-order members with user-controllable high-frequency numerical dissipation and adjustable sub-step locations, and $(s+1)$th-order members obtained by selecting the sub-step locations, with fixed dissipation. The latter reach up to seventh-order accuracy without increasing the number of sub-steps, although some high-order members are $A(\alpha)$-stable with stability angles extremely close to $90^\circ$. Analytical amplitude and phase errors further reveal parity-dependent superconvergence in undamped systems, and appropriate parameter selections can substantially increase either phase or amplitude accuracy beyond the formal order. Numerical benchmarks confirm the predicted convergence orders and the controllable suppression of spurious high-frequency responses.

Graph-based retrieval can improve multi-hop question answering, but existing approaches often incur high query-time costs and produce diffuse, oversized contexts that reduce generation efficiency. We present LiteRAG, a graph-based retrieval method that replaces expensive retrieval-time LLM control with query-conditioned algorithmic exploration and reasoning-chain context construction. On DistComp, a benchmark for multi-hop retrieval over distributed-systems papers, LiteRAG attains the highest overall quality among the evaluated methods (0.798) while reducing per-query latency by over 100$\times$ and cost by over 99% relative to GraphRAG Global and DRIFT. On UltraDomain, it matches LinearRAG on overall quality while using about 14$\times$ fewer tokens. An ablation study indicates that LiteRAG's query-adaptive thresholding and community-aware hub penalization are the main drivers of its token-efficiency gains.

A graph retrieval-augmented generation pipeline chooses which triples to put in the prompt, a syntax to write them in, an order to write them in, and a sentence telling the model what to do with them. We vary all four over six large language models and two knowledge-graph question answering benchmarks. Two of the four choices move the answer and the other two are flat. The first is whether the answer path, the triples needed to reach the answer, is in the prompt at all. Holding the number of triples fixed and replacing every triple that is not on the chain with material from an unrelated entity changes answer accuracy by +0.003 F1, while removing the chain costs most of what the graph was worth. Retrieval budget belongs on recall, and precision in the range we can test buys nothing. There is no retriever here: subgraphs come from gold SPARQL, so precision describes the context we build, not a system setting. The second is the grounding instruction. With no facts in the prompt, telling a model to answer using only the provided facts drops F1 from 0.299 to 0.035, a factor of 8.63. That figure describes an evaluation with an empty context arm rather than a working pipeline, and an experiment that applies the instruction to its context arm but not to its no-context baseline manufactures a spurious finding that graph context hurts at depth. We found one in our own results and retract it. Syntax, triple order and subgraph size produce no effect we can measure at multi-hop depth. The comparison that would price the grounding instruction against correct context is not measurable with a format-sensitive scorer, because the instruction determines the response format; we report it as an open contrast rather than a number.

Exact algorithms for solving mixed-integer bilevel linear programs (MIBLPs) typically rely on sequences of lower and upper bounds that converge to the optimal value. These procedures are commonly initialized using the single-level relaxation (SLR), obtained by omitting the follower's optimality condition and solving the resulting single-level optimization problem. In this paper, we investigate whether, for broad classes of MIBLPs, the resulting standard bounds admit uniform improvements that can be computed within the same computational complexity regime. For pure continuous bilevel linear programs, we show that, unless $P = NP$, neither the SLR-based lower bound nor its associated upper bound can be uniformly improved in polynomial time, even for the class of min-max problems. We then extend this analysis to the class of pure integer min-max bilevel linear programs under the assumption that the polynomial hierarchy does not collapse. First, we show that the continuous relaxation of the SLR admits no uniform polynomial-time computable improvement. We then prove that neither the SLR itself nor its associated upper bound admits a uniform improvement by a polynomial-time algorithm with access to a mixed-integer linear programming (MILP) oracle. Importantly, this rules out uniform improvements by iterative MILP-based approaches, including cutting-plane-based and decomposition algorithms. Overall, our results demonstrate that the SLR-based bounds are, in a complexity-theoretic sense, unimprovable systematically within their natural computational regimes.

Communication delays induced by cyber attacks present a critical challenge to the safe operation of connected autonomous driving. This study investigates the use of intention sharing communication strategy to enhance the resilience of model predictive controllers under Denial-of-Service attacks. We employ a vehicle-in-the-loop testbed integrating a real drive-by-wire vehicle with a microscopic traffic simulator and vehicle-to-X communication infrastructure. We emulate Denial-of-service attacks that induce communication delays of up to two seconds. We evaluate three controller variants: baseline status-sharing, intention-sharing, and delay-aware intention-sharing control. Experimental results reveal that while baseline control suffers significant performance degradation and frequent collisions under adversarial delay, intention sharing eliminates collisions and maintains behavior near nominal levels for the tested scenarios. These findings demonstrate the practical potential of intention-sharing architectures for safeguarding connected vehicles against network-layer degradation.

Cylindrical watertight enclosures are widely used across various underwater systems, from unmanned underwater vehicles (UUVs), to remotely operated vehicles (ROVs), to various sensor platforms. However, electronics are typically built on rectangular PCBs arranged in horizontal stacks, which inefficiently occupy the circular cross-section volume that is critical for both payload capacity and buoyancy management. This paper presents CUB (Central Utility Board) and CougarTail, a general-purpose system designed to address this gap. CUB is a circular PCB sized for 4-inch-diameter enclosures that consolidates a Raspberry Pi Compute Module 5 (CM5) and an STM32 microcontroller, while also providing power management features and auxiliary connections. Mounted coaxially, CUB reduces the electronics stack of our CougUV from 200 mm of tube length and 709 g to 25 mm and 156 g, returning that length and mass budget to payload and buoyancy trim. CougarTail is an open-source companion sensor mast that houses a GPS antenna and two dual-band (2.4 and 5 GHz) omnidirectional PCB antennas. Both components are validated through bench testing and integration on a CougUV platform, our small open-sourced torpedo UUVs.

Large language models (LLMs) now show strong reasoning and coding abilities, raising the prospect that they could assist in developing and optimizing the very infrastructure that powers them. Existing benchmarks, however, mainly focus on isolated kernels, predefined operators, or pre-specified optimization targets, and cannot evaluate open-ended, long-horizon LLM infrastructure engineering. To address this gap, we present $\Phi$-Bench, a benchmark of 85 tasks grounded in real LLM training and inference codebases. The tasks cover nine categories of LLM infrastructure and come in three formats, ranging from kernel function completion to long-horizon implementation and repository-wide end-to-end optimization. They are built by a construction pipeline that turns real pull requests, issues, and repository code into executable tasks with validated tests, and every reward is gated by correctness and anti-hacking checks. Across eight frontier models, the strongest model, Claude Opus 5, scores only 36.53%, and every model performs worse on long-horizon, repository-level implementation than on kernel-level tasks. Trajectory analysis further suggests that stronger models treat optimization as controlled experimentation, validating ideas with cheap local runs and ruling out confounders before drawing conclusions. We will open-source $\Phi$-Bench to facilitate future research on LLM agents' ability to develop and optimize the LLM infrastructure that powers themselves.

Feature transformation improves predictive performance on tabular data by constructing informative abstractions from raw features. Recent generative approaches encode transformation knowledge into continuous embedding spaces for efficient exploration of candidate strategies, but face three key limitations: (1) overlooking hierarchical relationships between low-level features, operations, and high-level abstractions; (2) enforcing order-sensitive embeddings on inherently permutation-invariant transformation sequences, thereby introducing systematic bias; and (3) relying on gradient-based search, which is ill-suited to non-convex transformation spaces. We propose a framework with two complementary components. First, a permutation-invariant hierarchical module captures interactions across features, operations, and abstraction levels, with a self-attention pooling mechanism that maps semantically equivalent structures to consistent embeddings aligned with downstream performance. Second, a policy-guided multi-objective reinforcement learning strategy initializes the search from empirically strong seeds and jointly optimizes predictive accuracy and transformation efficiency. Extensive experiments on diverse tabular benchmarks demonstrate the effectiveness and robustness of our framework against strong baselines. Our code and data are publicly available at: https://github.com/RayLiu1103/PHER.

Vision-language models such as CLIP embed images and text in a shared space, where modality-specific distributions often remain separated. Existing accounts connect this modality gap to initialization, contrastive dynamics, and information imbalance, while its distributional and pairwise contributions to retrieval remain unresolved. We introduce UOT-Gap, a training-free variational diagnostic that models frozen image and text embeddings with unbalanced entropic optimal transport (UOT). The UOT optimum separates transport, coupling complexity, and marginal mass variation; a complementary pair-aware residual compares observed image-caption pairs with the UOT soft matching. On Flickr8K and COCO-1K with frozen CLIP, OpenCLIP, and SigLIP encoders, caption degradation reduces Flickr8K Recall@1 from 0.559 to 0.003. Across six dataset-model conditions, the pair-aware residual tracks retrieval degradation with mean absolute Spearman 0.973, compared with 0.392 for the mean gap. The association remains stable across five random COCO-1K subsets at $0.954\pm0.026$, with a minimum of 0.943. UOT barycentric updates reduce the transport objective while degrading retrieval, distinguishing geometric objective descent from task improvement. These results establish UOT-Gap as a diagnostic for caption quality, modality alignment, and retrieval robustness.