Computation and Language
Total Recall at What Cost? Benchmarking the Serving Cost of Agentic Memory Systems
Natchanon Pollertlam and Witchayut Kornsuwannawit
Long-running conversational
agents increasingly rely on a
memory system to avoid resending the whole conversation each turn, yet how much that costs to serve has received little systematic benchmarking. We compare three
memory systems (Mem0, Hindsight, and Mastra Observational
Memory) against two reference strategies – a fixed-size rolling window and resubmitting the full transcript – across two backbones and conversations of up to 400 turns, pairing every cost measurement with answer accuracy on 665 LoCoMo questions. First, a
memory system's serving cost cannot be predicted from conversation length and message size alone: a regression that tracks the two reference strategies closely misses the
memory systems by 18-69%, their cost driven instead by internal
memory behavior. Second, a break-even analysis shows that whether – and when – a
memory system becomes cheaper to serve than the full transcript is highly sensitive to the system and the backbone, from the first tens of turns for the cheapest to never within 400 turns for the most expensive. Third, no system wins on both axes: accuracy spans 21-54%, and the backbone choice drives cost as much as the
memory system does.
Read the full abstract ▾
Cryptography and Security
ToolHazard: Scaling Adversarial Environments for Security Evaluation and Alignment of LLM-based Agents
Yutao Mou, Pengfei Yang, Zhe Yin et al. (+6 more)
Large
language model (LLM)
agents integrated with external tools are vulnerable to indirect
prompt injections embedded in environmental states. However, existing studies largely rely on manually implemented or reused environments,
stochastic LLM-based tool simulation, and predefined injection locations, limiting scalable security research across broader domains. To bridge this gap, we propose **ToolHazard**, a scalable
adversarial environment synthesis framework that reduces human engineering and supports expansion with additional seed domains and compute. Through an Environment Simulator, an Attacker
Agent, and a User Simulator, ToolHazard synthesizes executable stateful environments, discovers viable injection points and generates environment-specific payloads, and constructs state-grounded long-horizon tasks. Based on ToolHazard, we build **ToolHazard-Bench** for stress-testing
agents under complex workflows and diverse environmental attacks. Experiments reveal substantial
agent vulnerabilities and show that injection timing and placement affect attack effectiveness. Moreover, ToolHazard-generated
alignment data improves security on both ToolHazard-Bench and AgentDojo while preserving benign task utility.
Read the full abstract ▾
All authors Yutao Mou, Pengfei Yang, Zhe Yin, Zhangchi Xue, Xiaotian Luan, Dingyao Yu, Tong Zhang, Shikun Zhang, Wei Ye
Robotics
D3D-GEN: Robot-Aware Domain-Grounded Interactive 3D World Generation for Social Robotics
Anh Duc Do, Volodymyr Shcherbyna, Tai Duc Nguyen et al. (+5 more)
Training and validation of
Embodied AI for social
navigation critically depends on realistic simulation environments, yet many current approaches fail to find a balance between realism and simulability. We propose D3D-GEN, a novel world generation system that combines a domain
agent with a
retrieval-augmented generation (RAG) pipeline grounded in that domain. Our system enables users to rapidly generate domain-grounded, fully interactive 3D worlds by automating both the collection of domain knowledge and the synthesis of realistic floorplans and object placements, without dependence on any fixed 3D model database. Given a domain description
prompt, the research
agent collects publicly accessible domain-specific data and constructs a persistent domain database. Using this database, our RAG pipeline generates plausible floorplans and object placements by dynamically querying a user-provided semantic database, which can be easily extended or modified. The output is a fully interactive 3D world loadable by the popular simulators Isaac Sim and Gazebo. With our approach, we have built databases for several common domains (indoor residential, hospital, office) and generated dozens of distinct, plausible simulation environments for each domain. We present D3D-GEN with a local web frontend that facilitates rapid, interactive world generation for
robot simulation.
Read the full abstract ▾
All authors Anh Duc Do, Volodymyr Shcherbyna, Tai Duc Nguyen, Spaarsh Thakkar, Zhengcheng Shen, Teham Buiyan, Archan Misra, Linh Kästner
Machine Learning
DCM Bandits: Multiplayer Information Asymmetric Cascading Bandits for Multiple Clicks
Andy Wang, Charlton Shih and William Chang
In this work, we extend the Dependent Click Model (DCM) Bandits to a multiplayer information-asymmetric setting, where multiple
agents interact with a shared ranked list and may observe multiple clicks per session, introducing new challenges for selection strategies. We study asymmetry in (1) actions and (2) rewards, providing sublinear regret guarantees for three settings where at least one asymmetry is present. Establishing matching information-theoretic lower bounds for these settings is left as an open problem. We further show that for small termination probabilities, the termination ranking need not be known, improving on prior single-
agent results. Experiments confirm that our
algorithms perform well across asymmetric environments and highlight the critical role of feedback structure, specifically the distinction between full versus first-click feedback, in coordinating exploration and minimizing regret.
Read the full abstract ▾
Human-Computer Interaction
ATOM: Geometry-Aware Microgesture towards Object-Agnostic Tangible Interaction
Yinqiao Wang, Hao Xu, Qixuan Liu et al. (+3 more)
This paper presents ATOM, an integrated framework towards agnostic and tangible object interactions with microgestures. Our goal is to support microgesture interactions across different everyday objects, with the capability to automatically leverage the
geometric affordance of each object. We formulate a fingertip-aware detection pipeline to leverage generative 2D and 3D models for geometry enhancement and refinement. We then introduce a usability-based method to prioritize the detected elements based on their ergonomic suitability for interactions. Building on this foundation, we further develop an AR system to transform everyday handheld objects into tangible user interfaces with 0D, 1D, and 2D microgesture interactions. Across transitions among everyday cooking objects of varying shapes and sizes, ATOM outperformed ablation baselines in task completion, usability (SUS), and workload (NASA-TLX). A further study with 10 objects demonstrates ATOM's generalizability across objects and grasps, highlighting its potential towards fluid, object-agnostic tangible interaction in real-world AR scenarios.
Read the full abstract ▾
All authors Yinqiao Wang, Hao Xu, Qixuan Liu, Shengdong Zhao, Pheng-Ann Heng, Chi-Wing Fu
Robotics
Enhancing Visual Domain Robustness in Behaviour Cloning via Saliency-Guided Augmentation
Zheyu Zhuang, Ruiyu Wang, Nils Ingelhag et al. (+2 more)
In
vision-based behavior cloning (BC), conventional image augmentations such as Random Crop and Color Jitter often fall short under substantial visual domain shifts, including changes in shadows, distractors, and backgrounds. Superimposition-based augmentations, which blend in-domain and out-of-domain images, have shown promise for improving generalization in computer
vision, but their suitability for BC remains uncertain because task-critical semantics, spatiotemporal relationships, and
agent-target interactions must be preserved. To address this, we introduce RoboSaGA, a Saliency-Guided Augmentation method within the superimposition family tailored for
vision-based BC. RoboSaGA dynamically adjusts augmentation intensity at the pixel level using policy-driven saliency, enabling aggressive augmentation in task-irrelevant regions while preserving task-critical information. It integrates seamlessly into existing architectures without requiring structural modifications or additional learning objectives. Experiments in both simulated and real-world settings show that RoboSaGA preserves in-domain performance while substantially improving robustness to visual domain shifts, including distractor and background changes, as well as lighting and shadow variations. Code is available at https://github.com/Zheyu-Zhuang/RoboSaGA.
Read the full abstract ▾
All authors Zheyu Zhuang, Ruiyu Wang, Nils Ingelhag, Ville Kyrki, Danica Kragic
Machine Learning
Forward and Inverse Virtual Metrology for Phototransistor Gain: A Hierarchical, Uncertainty-Aware Approach for Small Production Datasets
Mahshid Amirabgir, Lorenza Ferrario, Paolo Conci et al. (+2 more)
The customization, optimization and stabilization of the process flow of a silicon bipolar phototransistor commits months of cleanroom time before a finished device can be measured, so a model that predicts device gain from process parameters before a run has value out of proportion to its accuracy. We study this problem on a real fabrication history, thirteen to fourteen process runs of a single device: a small-sample, hierarchically structured setting unlike the large-corpus regime of conventional virtual metrology. Decomposing the variance of device gain, we find that roughly half of it lies between process runs rather than within them, so recipe-only prediction is bounded by construction. Building on these findings we provide a forward gain predictor with a relative, uncertainty-aware signal, an inverse search that returns recipes for a target gain, and, as the foundation for all of it, a multi-level data-quality assessment tailored to the nested physical entities of fabrication (batch, wafer, die) with an explicit cross-level linkage score. The normalized
dataset and analysis code are released for full reproducibility.
Read the full abstract ▾
All authors Mahshid Amirabgir, Lorenza Ferrario, Paolo Conci, Mahdieh Amirabgir, Giancarlo Orengo
Neural and Evolutionary Computing
Lapis: Laplacian Spiking Attention via First-Spike Timing and Membrane Leakage
Kaiwen Tang, Jiaqi Zheng, Zixuan Zhu et al. (+3 more)
Self-
attention has become central to spiking
vision transformers, yet its query-key scoring is still largely inherited from dense networks. Existing spiking variants either simplify dot product scoring or replace it with discrete operators, but spike timing, the native variable of a spiking network, does not directly define how
tokens are related. We propose Lapis, a spiking
attention mechanism that scores each
token pair by the L1 distance between its query and key first-spike latency vectors under time-to-first-spike coding, and maps this distance to an affinity through a
Laplacian kernel. The kernel's exponential decay matches the impulse response of a leaky integrate-and-fire membrane, so the accumulated latency difference determines the decay of a membrane trace, while row normalization reduces to a bit shift under power-of-two rounding. Scoring therefore needs only subtraction, absolute value, and accumulation, and removes all multiplication between query and key channels. Under a matched backbone and training schedule, Lapis reaches 96.56% top-1 accuracy on CIFAR-10, within 0.53 points of dot-product scoring. On ImageNet-1K, it reduces the estimated arithmetic energy of the
attention path by 14.5x relative to dense dot-product
attention. The deployed 6-bit model attains 83.25% top-1 accuracy at an estimated arithmetic energy of 3.28mJ per image.
Read the full abstract ▾
All authors Kaiwen Tang, Jiaqi Zheng, Zixuan Zhu, Yiqun Wang, Zhanglu Yan, Weng-Fai Wong
Artificial Intelligence · Cross-listed from physics.optics
Two-Stage Deformable-Convolutional Inverse Design of Nanophotonic Absorbers from Optical Spectra
Waleed Waseer, Muhammad Shahid Jabbar, Muhammad Sohail Ibrahim et al. (+1 more)
Data-driven inverse design enables efficient generation of nanophotonic structures with prescribed optical responses, but spectrum-to-geometry mapping remains challenging due to non-uniqueness and fine
geometric features. This work presents a two-stage deformable-convolutional framework for reconstructing metal–
insulator–metal resonator geometries from 80-dimensional absorption spectra. The spectrum is projected to a $150\times4\times4$ latent representation and decoded into a $64\times64$ resonator mask. Training combines supervised reconstruction with least-squares
adversarial refinement initialized from the best supervised checkpoint. A three-run ablation compares deformable convolution with plain convolution, involution, Dynamic Conv, and ODConv under the same architecture. The proposed model achieves $20.79\pm0.31$~dB PSNR and $0.8501\pm0.0082$ SSIM, improving over plain convolution by 2.16~dB and 0.0831, respectively. It further achieves Dice $0.9623\pm0.0027$, IoU $0.9342\pm0.0038$, and boundary F-score $0.9550\pm0.0027$. Spectral consistency evaluated using a frozen forward surrogate yields RMSE $0.0805\pm0.0013$ and $R^2=0.7923\pm0.0065$. Learned offsets show stronger adaptive sampling at coarse and intermediate decoder stages. Overall, deformable sampling with supervised initialization and
adversarial refinement improves spectrum-conditioned geometry reconstruction.
Read the full abstract ▾
All authors Waleed Waseer, Muhammad Shahid Jabbar, Muhammad Sohail Ibrahim, Shujaat Khan
Machine Learning
Small-Scale Experiments: Are We There Yet?
Nicholas Lourie, Kyunghyun Cho, Karen Ullrich et al. (+1 more)
Scaling laws promised cost-effective experiments; six years later, they have yet to fully deliver. Instead, researchers have found them unreliable at small scales (starting at 4M parameters) and concluded that sizable models cannot be avoided. We show this is not the case: the confounding factor is hyperparameters. Small models are highly sensitive, but hyperparameter sensitivity fades with scale. This small-scale sensitivity makes scaling laws easy to miss because they only emerge on the fully tuned frontier, and reaching that frontier requires an extensive search far beyond what most ever run. By ablating the basic scaling law recipe, we show well-tuned hyperparameters matter more than any other ingredient. Further, we reveal why those hyperparameters become easier to find: as scale increases, the hyperparameter loss surface becomes lower dimensional. Nevertheless while scaling laws exist in small models, extrapolation hits statistical limitations. A holistic approach is required. Synthesizing our insights with the recent literature, we develop a new methodology for model-centric research and demonstrate it on a question that once took the field years to settle: where to place normalization layers in the
transformer architecture. From small-scale experiments, we recover the large scale result: pre-normalization works better as models grow in size. With the right tools and a better understanding, small-scale experiments can deliver on scaling laws' long-awaited promise.
Read the full abstract ▾
All authors Nicholas Lourie, Kyunghyun Cho, Karen Ullrich, Sanae Lotfi
Data Structures and Algorithms
A Tight Scale-Locality Bound for Partial Detection in Non-Adaptive Group Testing
Nader H. Bshouty
We give a lower bound for randomized non-adaptive group testing when the goal is to find any $\ell$ defective items but the total number $d$ of defectives is unknown. Bshouty and Haddad-Zaknoon proved an upper bound of $O(\ell\log^2 n)$ tests and a lower bound of $$\Omega\!\left(\frac{\ell\log^2 n}{\log \ell+\log\log n}\right).$$ We prove the matching lower bound. More generally, we show that every randomized non-adaptive
algorithm that succeeds with constant probability for every defective set must use $$\Omega\!\left(\ell\log^2(n/\ell)\right)$$ tests. The proof is as follows. At a fixed value of $d$, finding $\ell$ defectives requires about $\ell\log(n/d)$ bits of information. On the other hand, one fixed group test is informative only when its size is tuned to the scale of $d$; across all logarithmic scales of $d$, a single test contributes only $O(1)$ bits. Summing over all scales gives the lower bound. We also record the matching upper bound $$O\!\left(\ell\log^2(n/\ell)\right),$$ obtained by running the known-$d$
algorithm in parallel over dyadic guesses for $d$. Thus the randomized non-adaptive complexity of unknown-$d$ partial detection is $\Theta\!\left(\ell\log^2(n/\ell)\right)$ for constant success probability.
Read the full abstract ▾
Systems and Control
Disturbance-Observer-Based Grid-Forming Control for Unbalanced Grids
Juho Määttä, Marko Hinkkanen, Tuure Nurminen et al. (+4 more)
This article proposes a grid-forming control method for operation under unbalanced grid-voltage conditions. The method regulates the positive-sequence active power delivered to the grid and actively suppresses the negative-sequence converter voltage, thereby maintaining a constant converter-voltage magnitude during unbalanced faults when the converter operates within its physical limits. Only the converter current is measured on the AC side, and a disturbance observer is used to synchronize with the grid and to provide integral and resonant action. Estimates for the positive- and negative-sequence grid voltage are obtained from the disturbance observer. A current-limitation scheme for both balanced and unbalanced faults is integrated. Comprehensive stability analysis and tuning guidelines are provided. Experimental results using a 12.5-kVA converter demonstrate that the proposed method can operate during severe balanced and unbalanced faults.
Read the full abstract ▾
All authors Juho Määttä, Marko Hinkkanen, Tuure Nurminen, Orcun Karaca, Rayane Mourouvin, Jarno Kukkola, Lennart Harnefors
Information Theory · Cross-listed from quant-ph
Uniformly Stable Minimal Weyl–Heisenberg Measurements Approaching the SIC Benchmark
Xiuwu Zhu and Yu Wang
Informational completeness (IC) guarantees that an inverse exists, not that it is statistically well conditioned. For minimal rank-one Weyl–Heisenberg (WH) measurements, covariance makes the nonidentity projector-Gram spectrum proportional to the fiducial's ambiguity intensities, with
eigenvalues \(d|\chi_\phi(u)|^2\), turning stability into an explicit worst-direction design problem; write \(\lambda\) for its smallest nonidentity
eigenvalue. Haar fiducials are IC almost surely while \(\mathbb E[\lambda^{-1}]=\infty\), and an explicit
geometric family used to establish balanced informationally complete measurements in every dimension has a normalized spectral floor bounded above by an exponentially decaying envelope. We then construct a hierarchy of minimal measurements. A cyclic family with exactly \(d^2\) outcomes in every integer dimension has floors \(\Theta(d^{-3})\) for odd \(d\) and \(\Theta(d^{-5})\) for even \(d\); a finite-field family for \(q=2^m\) obeys the uniform bound \(\lambda\ge4/9\). Our main result treats every
prime-power dimension of characteristic \(p\ge5\). A balanced one-coordinate perturbation repairs the zero ambiguity axis of a cubic Alltop state, gives an attained floor uniformly bounded below by a positive constant, and confines the entire nonidentity spectrum to \([L_q,U_q]\) with \(U_q/L_q\to1\). Its SIC-normalized minimum tends to one, and \(\lambda(\phi_q)/\Lambda_q^\star\to1\) for the global finite-field WH max–min optimum \(\Lambda_q^\star\), without assuming SIC existence. The complete spectrum determines the exact finite-sample
Hilbert–Schmidt error of canonical linear inversion at \(I/d\), while its lower edge controls local Fisher efficiency and canonical-shadow bounds.
Read the full abstract ▾
Numerical Analysis · Cross-listed from math-ph
A uniform elliptic reduction, an order-matching criterion, and precision benchmarks for the strong-coupling Birman-Schwinger analysis of the lattice three-boson trimer
Mikhail Dolgopolov
We present a rigorous strong-coupling Birman-Schwinger analysis of the three-
boson lattice Schroedinger operator on Z^2 at the exceptional quasimomentum K = pi. First, we provide an exact closed-form
benchmark for the fiber Fredholm determinant, valid for every quasimomentum K, obtained via a uniform
elliptic reduction. Second, we formulate and prove a general order-matching criterion determining whether a leading-order Fredholm determinant asymptotic, with relative error O(1/mu), suffices to fix the constant-order additive energy correction, or if the next-order refinement is required. Applying the criterion to the formal branch z = -2mu + d, we identify an algebraic crossing -2mu + 6 + 8/mu + O(mu^{-2}), but demonstrate that this crossing does not correspond to a true
eigenvalue of the full
Hamiltonian. The actual ground state obeys the rigorous variational bounds -3mu <= z_1^{pi,s}(mu) <= -3mu + 6, so that z_1^{pi,s}(mu) = -3mu + O(1), the same leading branch as at K = 0; direct finite-volume diagonalisation confirms the refined asymptotic -3mu + 6 + O(mu^{-1}) and spectral gap 2mu - 2 + O(mu^{-1}). We independently confirm that the known K = 0 constant C approximately 3.96458 requires no analogous refinement. Finally, we compare the asymptotic precision levels achieved across recent
lattice few-body models and draw a structural parallel with the parity-based classification of topological band
insulators at time-reversal-invariant momenta.
Read the full abstract ▾
Computer Vision
LookBack: Where and How to Score LVLM Responses via Visual Reference Usage
Beomsik Cho, Jinhyeong Kim, Dongseok Lee et al. (+1 more)
Large
Vision-
Language Models (LVLMs) integrate visual perception with
language generation, enabling responses that span image understanding and complex
reasoning. However, LVLMs do not just inherit the text-level
hallucinations; they also hallucinate against the image, producing fluent responses ungrounded in what they see. This makes LVLM response scoring inherently harder, and our diagnostics show that existing confidence-based metrics adopted from LLMs are insufficient for LVLMs. Specifically, removing the input image barely changes confidence-based selection, suggesting that output-space confidence primarily captures textual plausibility rather than agreement with the image. To address this gap, we propose LookBack, a training-free LVLM response scoring method that augments
token likelihood with visual lookback score, a lightweight measure of how strongly each response
token refers to image
tokens. Across four
benchmarks and three models, LookBack consistently improves Best-of-$N$ selection over existing baselines with negligible additional overhead.
Read the full abstract ▾
All authors Beomsik Cho, Jinhyeong Kim, Dongseok Lee, Jaehyung Kim
Information Retrieval
From Overlooked to Explored: Recovering Item Relations via Mixture of Perspectives for Sequential Recommendation
Junyoung Kim, Wonbin Kweon, Woojoo Kim et al. (+3 more)
Capturing user preference from a user's interaction sequence is the central challenge of Sequential Recommendation (SR). This preference intuitively emerges from inter-item relations: each item transition reflects a preference embedded in the relations between items, making the faithful capture of these relations essential for accurate recommendation. For this reason, self-
attention is dominant in sequential recommendation for its ability to compute pairwise item interactions, yet our empirical analysis reveals that it consistently suffers from similarity bias across various types of
transformer-based SR models: dot-product
attention scores disproportionately favor similar items, systematically overlooking heterogeneous relations with meaningful preference signals and directly limiting recommendation performance. To address this, we propose PRISM (Perspective-based Relational Insight Synthesis Module), a module that re-examines item relations from multiple perspectives. PRISM employs K Perspective Lenses to calibrate
attention from distinct viewpoints, combining an Affinity View that refines homogeneous relations and a Contrast View that exposes heterogeneous ones suppressed by similarity bias, enabling the model to capture the full spectrum of user preferences. Extensive experiments on seven real-world
benchmarks demonstrate that PRISM consistently outperforms state-of-the-art baselines. Our code is available at https://github.com/327aem/PRISM/.
Read the full abstract ▾
All authors Junyoung Kim, Wonbin Kweon, Woojoo Kim, Jaehyung Lim, Dongha Kim, Hwanjo Yu
Multimedia · Cross-listed from eess.IV
ResPCC: A Loss-Resilient Neural Point Cloud Codec over Lossy Networks
Xueqin Niu, Mufan Liu, Yifan Wang et al. (+3 more)
Point cloud compression (PCC) is critical for efficient storage and transmission of 3D data. While recent learning-based PCC methods achieve good rate-distortion (R-D) performance, they generally rely on ideal transmission conditions. In practice, packet loss is a common issue and can severely distort latent features, causing coordinate drift and
geometric degradation. To address this challenge, we present ResPCC, the first end-to-end neural point cloud codec designed to offer intrinsic resilience against data loss. Our framework is loss-rate-aware and adapts to diverse packet loss conditions. At the encoder, we introduce a Condition-Adaptive Latent Modulation (CALM) module to adjust latent feature distributions according to the perceived loss rate, as well as a Spatial-Channel Interleaving (SCI) mechanism that transforms channel-wise data extinction into spatially scattered element-wise missing patterns. At the decoder, we develop a Mask-Aware Graph-based Latent Restoration (MGLR) module, followed by a Dictionary-based Refinement (DBR) stage to recover corrupted features and align them with canonical priors. Evaluations on ShapeNet and SemanticKITTI under 5% to 30% packet loss rates show that ResPCC consistently delivers superior stability and R-D performance over baselines. Our framework maintains high reconstruction fidelity under lossy conditions, providing a reliable solution for 3D data transmission over practical networks. Code is available at https://github.com/starrynight314/ResPCC.
Read the full abstract ▾
All authors Xueqin Niu, Mufan Liu, Yifan Wang, Le Yang, Jun Sun, Yiling Xu
Computer Vision
BoltNet: An Ultra-Lightweight Convolutional Network for On-Device Plant Species Identification
Daniel Rossi, Guido Borghi and Roberto Vezzani
Automated plant species identification from citizen-science imagery is an established, demanding fine-grained recognition problem: large taxonomic label spaces, visually similar species, and long-tailed observations require real model capacity, while field use constrains
memory, latency, and power. Model size is only part of the deployment cost: intermediate activations held in
memory during inference and platformdependent execution behavior matter too, so compact recognition must be assessed on target hardware rather than through complexity metrics alone. We present BoltNet, an ultra-lightweight fully convolutional architecture combining a Spatial Redistribution Bottleneck and Logit PreSampling to improve the tradeoff between predictive performance and model size in high-cardinality classification, and report the AccuracyCompression Tradeoff as a complementary diagnostic. On Pl@ntNet300K, BoltNet reaches 0.682 F1-score with 341K parameters (1.37 MB), the highest F1-score among evaluated models below 2 MB and close to substantially larger convolutional backbones. Model-only measurements on a Raspberry Pi 5, Jetson Orin Nano, and Hailo-8 characterize execution across CPU, GPU, and NPU platforms, where BoltNet is the most consistently efficient model, with the best FPS/W on the GPU and NPU and second-best on the CPU. Results on AIDERv2 and CLRS provide secondary evidence of transfer across environmental image-classification tasks. Code available at: https://codeberg.org/danielrossi/BoltNet
Read the full abstract ▾
Computation and Language
When the Knowledge Base Becomes the Gold Standard: Measuring Resource-Shared Evaluation Loops in Entity-Level Machine Translation
Jinhyung Bae, Dain Kil, Seongmin Oh et al. (+1 more)
The Seungjeongwon Ilgi, a UNESCO
Memory of the World record, is only 37.4% translated, and the most conspicuous failure mode in automatic translation is the person name – a misread name corrupts the historical fact rather than merely the surface. Low-resource historical domains have no expert gold standard for entity translation, so practitioners substitute a knowledge base (KB) for the gold. That KB is the same resource injected into the system: scoring becomes self-referential and the metric measures instruction compliance rather than translation quality. We measure this loop. Using expert person-name annotations from the National Institute of Korean History as a gold independent of the injection pipeline, we hold the entity set fixed and vary only the provenance of the correct reading. Of 527 expert-annotated mentions, only 31.1% lie outside the injection pipeline, and the residual loop is not uniform – in the overlapping segment the injected reading agrees with the human translation 97.8% of the time against 70.1% in the independent one, so the segment that looks healthiest is the one the loop is holding up. Across four models, a difference-in-differences analysis shows the gain from KB injection is confined to the segment whose gold shares the injected resource; in the independent segment it is at or below zero. Post-injection preservation clusters in a narrow 0.910-0.996 band even though baseline capability differs fivefold, so the reported gain is the complement of prior performance and weaker models appear to improve more dramatically. On an independent sample built by removing the construction filter, the measure replicates within model (overlapping intervals) while discriminating between models (non-overlapping intervals) – it reflects a property of the model, not of the sample.
Read the full abstract ▾
All authors Jinhyung Bae, Dain Kil, Seongmin Oh, Seungmin Lee
Distributed Computing
User-Assisted Collaborative Distributed Inference for Efficient QoS-Aware Autoscaling
Alfreds Lapkovskis, Ali Beikmohammadi, Sindri Magnússon et al. (+1 more)
Growing demand for artificial intelligence (AI) inference services requires scalable infrastructure, yet centralized serving costs rise with demand. We propose a collaborative distributed inference system combining dedicated infrastructure with resources contributed by service users. Dedicated resources provide baseline capacity for maintaining quality of service (QoS), while volunteered resources absorb increasing demand without proportional growth in centralized infrastructure. To capture
stochastic and dynamic interactions among users, resources, tasks, and policies, we develop a high-dimensional generative Markov model with structured temporal factorization. The model supports simulation and provides a foundation for task scheduling and QoS-aware resource allocation optimization. We evaluate the system across user populations, resource capacities, and centralized and distributed scheduling policies. Simulations show that distributed scheduling becomes increasingly advantageous as the user population grows, improving request completion and P99 latency while substantially reducing dedicated resource consumption. These results demonstrate the feasibility of user-assisted collaborative inference for infrastructure-efficient autoscaling.
Read the full abstract ▾
All authors Alfreds Lapkovskis, Ali Beikmohammadi, Sindri Magnússon, Praveen Kumar Donta
Machine Learning
Air Quality Station Simulation via LSTM and Attention-Based Modelling
Alexander Kostadinov, Petar O. Hristov and Dessislava Petrova-Antonova
Poor air quality in urban areas is driven by a complex chain of processes and presents a significant public health concern. To better understand and control the mechanisms that determine air quality, cities deploy networks of measurement stations, and launch initiatives for collecting denser data about the concentration of pollutants in the atmosphere. Extracting insights from the stations relies on their reliable and uninterrupted operation. However, hardware is susceptible to faults and black- outs that may result in data unavailability, which affects the overall quality of analyses. In this paper, we present a deep-learning model, called SATADL, which can infer complex relations and output multiple-hour-ahead air-quality forecasts. The goal of the model is to simulate the mea- surements of an unresponsive station until its operation is restored. The architecture of the model, which allows it to extract information from different aspects of the data, is described in detail and a careful examination of all of its components is provided. We demonstrate the performance of SATADL on four sets of air quality stations from around the world, by using it to simulate the concentration of PM10 for periods of hypothetical failures of one of the measurement stations, lasting for as long as 48 hours. A selection of baseline and published deep learning models were trained and used as a
benchmark. The results show that SATADL per- forms better across different prediction windows, for both coefficient of determination and root mean squared error, demonstrating its suitability as a virtual proxy station.
Read the full abstract ▾
Computer Vision
GeoBridge: Decoupled Semantic Conditioning for Generative Image Geolocalization
Zhiyang Dou, Xumeng Han, Fengde Peng et al. (+4 more)
Multimodal large
language models (MLLMs) have advanced image geolocalization mainly by improving how they reason about geographic cues. How that
reasoning isdecoded into coordinates, however, has lagged behind. Predicting a place name for a geocoding API is discrete and lossy: it ignores image evidence and collapses multi-granular semantics into a coarse lookup. We argue that the bottleneck has shifted from what a model reasons to how that
reasoning is represented for a continuous, geometry-aware decoder. We present GeoBridge, a role-decoupled conditioning mechanism that connects a frozen semantic MLLM to a frozen Riemannian flow-matching head that generates coordinates on the sphere. The central obstacle is arole conflict: supervising the condition with discrete semantic labels biases its representation toward class-discriminative geometry, at odds with the smooth
manifold the generative head requires. GeoBridge keeps the semantic supervision decoupled from the condition interface: a separate projection forms the continuous condition the frozen head expects, injecting geographic priors without disturbing the spherical decoder. On IM2GPS3K, GeoBridge reaches 38.67/52.89/70.37 at the 25/200/750 km thresholds, improving over a place-name-to-API pipeline and
reasoning-augmented direct prediction at these precision-relevant scales. GeoBridge is a decode-side algorithmic contribution, orthogonal and complementary to chain-of-thought
reasoning. Code will be made publicly available.
Read the full abstract ▾
All authors Zhiyang Dou, Xumeng Han, Fengde Peng, Zipeng Wang, Moxuan Zhao, Zhipei Huang, Zhenjun Han
Databases
Enabling Differentiated QoS Degradation for Replicated Databases under Failures
Belkis Djeffal, Pierre Bourhis and Romain Rouvoy
Elasticity is commonly presented as the default response to capacity loss after failures, since replacement replicas can compensate for failed nodes and restore pre-incident service levels. Replacement capacity entails both delay and additional resource commitment, as replicas must be provisioned and synchronized before they can serve traffic. Under fixed budgets or constrained operating conditions, capacity restoration cannot be treated as the immediate recovery path. Failure handling must instead define how the service continues while capacity remains reduced. When the service exposes differentiated service levels, capacity loss cannot be handled uniformly. Degradation becomes part of the service behavior, requiring explicit control over how reduced capacity affects each class without erasing the intended differentiation. We study differentiated QoS degradation in replicated database services with service-class-aware sessions. We present a repair-to-target policy, implemented in PLB, a PostgreSQL JDBC middleware load balancer for service-class-aware routing. When fail-stop failures remove part of the available capacity, PLB updates the role assignment of healthy replicas into Premium, Mixed, and Freemium roles. This keeps the replica pool shared while ensuring that new session assignments continue to reflect the service class. We evaluate PLB under single and cascading replica failures across two deployment strategies: isolated perclass replica pools and shared, priority-agnostic routing. The results show that PLB improves median Premium goodput retention by 26-28 percentage points under a Premium-side fault, achieves more than 2x higher Premium goodput in the most severe cascading-failure phase, and reduces Premium p95 latency by 18.2% relative to shared round-robin.
Read the full abstract ▾
Computer Vision
Andreas Robinson, Abdelrahman Eldesokey and Michael Felsberg
Semi-supervised
video object
segmentation is a challenging task that aims to segment a target throughout a
video sequence given an initial mask at the first frame. Discriminative approaches have demonstrated competitive performance on this task at a sensible complexity. These approaches typically formulate the problem as a one-versus-one classification between the target and the background. However, in reality, a
video sequence usually encompasses a target, background, and possibly other distracting objects. Those objects increase the risk of introducing false positives, especially if they share visual similarities with the target. Therefore, it is more effective to separate distractors from the background, and handle them independently. We propose a one-versus-many scheme to address this situation by separating distractors into their own class. This separation allows imposing special
attention to challenging regions that are most likely to degrade the performance. We demonstrate the prominence of this formulation by modifying the learning-what-to-learn (LWL) method to be distractor-aware. Our proposed approach sets a new state-of-the-art on the DAVIS 2017 val
dataset, and improves over the baseline on the DAVIS 2017 test-dev
benchmark by 4.6 percentage points.
Read the full abstract ▾
Programming Languages
Inferring Empirical Sound Resource Bounds via Symbolic Execution and Linear Programming (Extended Version)
Samuel Frontull, Manuel Meitinger and Georg Moser
Existing approaches to resource analysis of programs can be classified into two main paradigms: static analysis and dynamic analysis methods. The former allow for formal guarantees but are inherently incomplete; the latter are widely applicable but may miss rare but characteristic (worst-case) scenarios and thus lack soundness. Hybrid approaches attempt to combine the strengths of both paradigms, thereby enabling the analysis of programs that are either too complex for purely static techniques or where dynamic approaches suffer from combinatorial explosion. In this paper, we present a novel hybrid approach that systematically derives upper bounds for the worst-case resource consumption of functional programs. Our method combines dynamic symbolic execution to exhaustively explore all possible computation paths within a constrained input space with mixed-integer linear programming to derive empirically sound upper bounds. We have implemented the methodology in a prototype tool, dubbed CompAS, which we made available on Zenodo.
Read the full abstract ▾
Machine Learning
Kernel Methods for Learning Operators with Multiple Inputs and Outputs
Adrien Weihs, Chunyang Liao, Jingmin Sun et al. (+1 more)
Learning mappings between infinite-dimensional objects is a central challenge in scientific machine learning. We introduce a general kernel-based encoder-decoder framework for operator learning that separates observation, representation, learning, and reconstruction. We develop this framework for multi-input, multi-output operator learning, where operators map between products of potentially distinct function spaces. Our approximation theory shows that, although the number of inputs and outputs can increase, the convergence rate is governed by the most challenging constituent approximation problem rather than the overall problem dimension. The framework leads to practical kernel methods with closed-form training and inference, combining mathematical tractability with computational efficiency. We further specialize the approach to multiple operator learning by introducing KernelMO, a family of kernel methods with complementary operator-valued and product-space formulations. Across five families of parametric partial differential equations, the proposed methods achieve competitive or state-of-the-art predictive accuracy while reducing training and inference costs relative to neural operator architectures and deep learning based models, offering an efficient and lightweight alternative.
Read the full abstract ▾
All authors Adrien Weihs, Chunyang Liao, Jingmin Sun, Hayden Schaeffer
Computers and Society
Quantifying the Relationship Between Clinical Safety and Environmental Impact in Therapeutic LLMs
Alireza A. Safaei, Laura M. Vowels, Matthew J. Vowels et al. (+2 more)
The deployment of large
language models (LLMs) in mental health contexts raises questions about the relationship between clinical safety and environmental cost. In this paper, we examine this relationship by combining K-Bench clinical safety scores with EcoLogits life-cycle assessment estimates across 47 supported model configurations. We evaluate model performance and environmental impact across four dimensions: energy use, carbon emissions, water consumption, and abiotic depletion. The results indicate a non-linear trade-off at the upper end of the safety distribution: a 2.61 percentage-point increase in clinical safety score corresponded to an approximately 60-fold increase in estimated energy use per million output
tokens. Row-level analyses further suggest that additional test-time compute did not consistently improve clinical safety and, in some configurations, was associated with lower clinical safety scores. These findings suggest that relying solely on larger models or additional inference-time computation may be an inefficient strategy for improving safety in therapeutic AI systems. We discuss the implications for sustainable deployment and highlight dynamic model selection, including model cascading, as a potential approach for reducing environmental impact while preserving clinical performance in higher-risk cases.
Read the full abstract ▾
All authors Alireza A. Safaei, Laura M. Vowels, Matthew J. Vowels, Apoorv Jha, Shekoufeh Rahimi
Machine Learning
Towards Understanding On-Policy Distillation through the Lens of Test-Time Scaling
Xinmu Ge, Zizhuo Zhang, Yu Huang et al. (+8 more)
On-policy
distillation (OPD) has emerged as a promising post-training technique for enhancing LLM
reasoning. Under the reverse KL objective, the idealized optimum of OPD aligns the student distribution with that of the teacher. When the teacher consistently outperforms the student, this naturally suggests that OPD should yield broad improvements over the pre-OPD student. However, do such improvements extend across the entire range of test-time sampling budgets? In this work, we revisit this expectation through the lens of test-time scaling by varying the sampling budget $K$ and evaluating performance with pass@$K$. Across multiple settings, we observe two distinct patterns: OPD can improve pass@$K$ at both small and large sampling budgets, but it can also improve small-budget performance while reducing large-budget pass@$K$. We show one condition that guarantees such a reversal and an idealized reverse KL counterexample where it occurs even when the teacher has higher accuracy on every problem. To choose between two candidate teachers at a target sampling budget, we propose the \textit{Teacher Advantage Score at $K$} (TAS@$K$), which can be computed before OPD training to predict which teacher will lead to a larger improvement in pass@$K$. Across three domains and thirteen
benchmarks, the ordering predicted by TAS@$K$ agrees with the observed pass@$K$ improvements of the resulting OPD models in 83.6% of experiments, providing a useful signal for teacher selection at the target pass@$K$.
Read the full abstract ▾
All authors Xinmu Ge, Zizhuo Zhang, Yu Huang, Jianing Zhu, Lin Yuan, Wanli Gu, Weichang Wu, Weiran Huang, Bo Han, Xiaolu Zhang, Jiangchao Yao
Information Theory · Cross-listed from eess.SP
A Universal Random Precoding Framework for MIMO Systems
Jiazhen Dong, Lei Liu, Xiaojun Yuan et al. (+1 more)
Current wireless systems combat inter-symbol interference (ISI) by diagonalizing or sparsifying the channel
matrix, yet they remain vulnerable to selective fading. To address this, we propose a universal random precoding (RP) transmission framework based on the universality class. RP leverages random transforms to statistically exploit all subchannels and construct an equivalent channel belonging to the universality class, thereby enhancing diversity gain while maintaining backward compatibility with existing waveforms. Low-complexity implementations include the randomly permuted fast transform (FT-RP) and the interleaved block-sparse fast transform (IBSFT-RP). A cross-domain OAMP/MAMP (CD-OAMP/MAMP) detector is designed for RP systems, which is replica maximum a posteriori (MAP)-optimal according to state
evolution (SE). Simulation results on
MIMO systems demonstrate that RP with CD-OAMP/MAMP achieves near-RM performance with much lower complexity, with additional benefits of flexible compression ratios for spectral efficiency.
Read the full abstract ▾
All authors Jiazhen Dong, Lei Liu, Xiaojun Yuan, Baoming Bai
Computation and Language
Located but Not Releasable: Silent Gate Inversion and Bounded Linear Release
Xining Xun
A growing body of work reports that
language models represent task-relevant latent structure that they fail to use. Whether such structure, once located, can be converted into behavior is a separate question that is rarely tested end to end. We submit the complete pipeline – detect, localize, and release – to a fully preregistered stress test on a 25.7M
transformer trained on causal-evidence discrimination, where a known suppression phenomenon (latent causal structure present but behaviorally unused) has previously been documented. Every threshold, claim template, and decision-tree branch was hashed and archived before any corresponding data existed. Three findings. (i) Localization succeeds: interventions at observation-evidence channels of mid layers restore target behavior on otherwise-suppressed worlds (paired release advantages $0.563$ and $0.854$, 97.5% CIs excluding zero; best-site release rate $0.889$). (ii) Gating fails out of distribution: a detector calibrated to trigger on zero out-of-distribution calibration worlds triggers on 6.9-7.3% of held-out in-distribution generations and on zero of the 2,400 held-out generations that actually need it – a complete inversion that silently reduces the gated pipeline to its base model. (iii) Linear release is capped: removing the gate and injecting a per-instance linear direction unconditionally yields a monotone dose-response that plateaus far below the preregistered release margin (intercept $0.382 \to 0.311 \to 0.264$ vs. threshold $\le 0.08$); per-instance adaptivity adds less than $\pm 0.03$. The failure is doubly located: the detector is OOD-inverted, and the entire family of linear release directions at this site and resolution is bounded away from sufficiency. The two failures are dissociable, and neither overturns localization. Every number traces to a hashed artifact in the released audit chain.
Read the full abstract ▾