arXiv preprints from January 1, 2026 through July 21, 2026 — 17:24:40 EST
The fundamental premise of Vision-Language-Action (VLA) models is to harness the extensive general capabilities of pre-trained Vision-Language Models (VLMs) for generalized embodied intelligence. However, standard robotic fine-tuning inevitably disrupts the pre-trained feature space, leading to "catastrophic forgetting" that...
In this extended abstract, we propose a novel research topic in the field of agentic AI, which we refer to as self-coding information systems. These systems will be able to dynamically adapt their structure or behavior by evaluating potential adaptation decisions, generate source code, test, and (re)deploy their source code...
As mobile applications (apps) become ubiquitous in everyday life, it is crucial for developers to prioritize accessibility for users with diverse abilities. While previous research has identified widespread accessibility issues and raised awareness of developer challenges, there remains a lack of cross-platform, globally...
Neural image codecs achieve higher compression ratios than traditional hand-crafted methods such as PNG or JPEG-XL, but often incur substantial computational overhead, limiting their deployment on energy-constrained devices such as smartphones, cameras, and drones. We propose Grayscale Image Compression with Differentiable Logic...
The attraction term in an equation of state for gases, $-a c^2$, proposed by Rankine in 1854, is generally related to the London dispersion force via the equation for the second virial coefficient, $B_2$, given by $B_2 = 2πN_0 \int_0^\infty \left(1- \exp \left(-ω/ kT\right)\right) r^2 \text{d}r$, where $ω$ is the attraction energy...
In cloud block store, indexing is on the critical path of I/O operations and typically resides in memory. With the scaling of users and the emergence of denser storage media, the index has become a primary memory consumer, causing memory strain. Our extensive analysis of production traces reveals that write requests exhibit a strong...
Perception in granular media remains challenging due to unpredictable particle dynamics. To address this challenge, we present SandWorm, a biomimetic screw-actuated robot augmented by peristaltic motion to enhance locomotion, and SWTac, a novel event-based visuotactile sensor with an actively vibrated elastomer. The event camera is...
As Multimodal Large Language Models (MLLMs) acquire stronger reasoning capabilities to handle complex, multi-image instructions, this advancement may pose new safety risks. We study this problem by introducing MIR-SafetyBench, the first benchmark focused on multi-image reasoning safety, which consists of 2,676 instances across a...
We study the Universal Difference Property (UDP) introduced by Altınok, Anders, Arreola, Asencio, Ireland, Sarıoğlan, and Smith, focusing on the relationship between the structural properties of a graph and UDP. We present condtions for when UDP must hold on unicyclic graphs. We then prove that if UDP does not hold on an edge-labeled...
We show that given a log-singular circle homeomorphism $h$ and given any $s\in[1,2]$, there is a flexible curve of Hausdorff dimension $s$ with welding $h$. We also see that there is another curve with welding $h$ and positive area. In particular, this implies that given a flexible curve $Γ$, there is a homeomorphism of the plane...
Synthetic data offers a promising solution for mitigating data scarcity and demographic bias in mental health analysis, yet existing approaches largely rely on pretrained large language models (LLMs), which may suffer from limited output diversity and propagate biases inherited from their training data. In this work, we propose a...
We study how document chunking choices impact the reliability of Retrieval-Augmented Generation (RAG) systems in industry. While practice often relies on heuristics, our end-to-end evaluation on Natural Questions systematically varies chunking method (token, sentence, semantic, code), chunk size, overlap, and context length. We use a...
A sequence $S=(x_1,\ldots, x_k)$ in $\mathbb Z_p$ is called a $(Q_p,\mathbf 1)$-weighted zero-sum sequence if there exist $a_1,\ldots,a_k\in Q_p$ such that $a_1x_1+\cdots+a_kx_k=0$ and $a_1+\cdots+a_k=0$. The constant $E_{Q_p,\mathbf 1}$ is defined to be the smallest positive integer $k$ such that every sequence of length $k$ in...
Identifying when and where a news image was taken is crucial for journalists and forensic experts to produce credible stories and debunk misinformation. While many existing methods rely on reverse image search (RIS) engines, these tools often fail to return results, thereby limiting their practical applicability. In this work, we...
Perturbation screens hold the potential to systematically map regulatory processes at single-cell resolution, yet modeling and predicting transcriptome-wide responses to perturbations remains a major computational challenge. Existing methods often underperform simple baselines, fail to disentangle measurement noise from biological...
We study the set of chords of a real-valued continuous function on [0,1] with f(0)=f(1)=0. We describe which chords may appear as isolated points and provide examples illustrating our characterization. Maximal Hopf sets are introduced and analyzed.
We show that the large $n$ limit of the $O(n)$ quantum rotor model defined on a general graph has the same critical behavior as the corresponding quantum spherical model and that the critical exponents depend solely on the spectral dimension $d_s$ of the graph. To this end, we employ a classical to quantum mapping and use known...
In this article we give general neccessary and sufficient conditions to ensure that a pseudo-Riemannian manifold is conformal to an Einstein space. These conditions are algorithmic in \emph{the metric tensor} whenever the Weyl endomorphism is invertible. Our conditions depend in an essential manner on the $\mathcal{C}$-connection. We...
How do authoritarian regimes strengthen global support for nondemocratic political systems? Roughly half of the users of the social media platform TikTok report getting news from social media influencers. Against this backdrop, authoritarian regimes have increasingly outsourced content creation to these influencers. To gain...
Scalar-rigid maps are Riemannian submersions by works of Llarull, Goette--Semmelmann, and the second named author. In this article we show that they are essentially Riemannian products of the base manifold with a Ricci-flat fiber. As an application we obtain a Llarull-type theorem for non-zero degree maps onto products of manifolds of...
We define the notion of torically hyperbolic varieties and we construct pair-of-pants decompositions for these in terms of angle sets of essential projective hyperplane complements. This construction generalizes the classical pair-of-pants decomposition for hyperbolic Riemann surfaces. In our first main theorem, we prove that the...
Liquid Time-Constant networks (LTCs), a type of continuous-time graph neural network, excel at modeling irregularly-sampled dynamics but are fundamentally confined to Euclidean space. This limitation introduces significant geometric distortion when representing real-world graphs with inherent non-Euclidean structures (e.g.,...
Kleene algebra (KA) is an important tool for reasoning about general program equivalences, with a decidable and complete equational theory. However, KA cannot always prove equivalences between specific programs. For this purpose, one adds hypotheses to KA that encode program-specific knowledge. Traditionally, a map on regular...
We investigate accretion disk dynamics and the formation of quasi-periodic oscillations (QPOs) in the infrared limit around Kerr-like black holes in asymptotically safe gravity. Relativistic hydrodynamic solutions of Bondi-Hoyle-Lyttleton (BHL) accretion reveal that quantum corrections significantly modify the structure of the shock...
Explainable AI (XAI) has become critical as transformer-based models are deployed in high-stakes applications including healthcare, legal systems, and financial services, where opacity hinders trust and accountability. Transformers self-attention mechanisms have proven valuable for model interpretability, with attention weights...