arXiv:2606.05129v1 Announce Type: new
Abstract: Preserving data privacy is an important topic in structural data management and data mining. However, the issue of privacy leakage in distributed causal structure learning is a persistent challenge, especially in cases where data transmission and computation are required. In this paper, we propose a method based on fully homomorphic encryption (FHE) that performs calculations on ciphertexts, keeping data encrypted in transition and computation. Nevertheless, adopting FHE to causal structure learning is challenging due to the high computation cost and limited support on division as well as logarithm operations in FHE. To tackle this challenge, we propose a series of novel techniques including (i) circuit simplification for better efficiency, (ii) approximation of division and logarithm through Newton-Raphson Reciprocal and Taylor expansion, and (iii) a batching technique with SIMD-acceleration to enhance the whole learning process. Additionally, our method can be easily extended beyond FHE by demonstration of its portability to support differential privacy. Empirical results show that our method achieves high consistency and comparable causal structure with the plaintext version in the datasets tested. Last, our method is efficient and practical to complete learning causal structures in tens of minutes even under the privacy protection of FHE.
Science Journals
arXiv:2606.05130v1 Announce Type: new
Abstract: Individual-level mobility prediction is central to urban simulation, transportation planning, and policy analysis. Supervised sequence models achieve strong accuracy but require task-specific training and offer limited decision-level transparency. Recent LLM-based methods improve interpretability, yet mostly rely on static prompts and single-pass inference, limiting their ability to seek additional evidence when mobility signals are weak or conflicting. We propose \method{}, a training-free LLM-driven agent framework that formulates next-location prediction as adaptive evidence-controlled decision making. \method{} resolves routine cases through a fast path based on historical regularity, while ambiguous cases trigger iterative tool use over recent trajectories, historical behavior, stay-move likelihood, and geographical evidence. Across three mobility datasets, AgentMob achieves the strongest overall performance among training-free LLM-based methods, with GPT-5.4 reaching 71.42\% Acc@1 on BW, 33.14\% on YJMob100K, and 33.50\% on Shanghai ISP. On BW non-fast-path cases, the LLM controller improves Acc@1 from 30.65\% to 48.62\% over a same-tool statistical baseline, showing that its main benefit lies in resolving ambiguous predictions through adaptive evidence gathering. Our code is available at https://github.com/Unknown-zoo/AgentMob.
arXiv:2606.05131v1 Announce Type: new
Abstract: Koopman theory turns nonlinear dynamics into a linear spectral problem. In computation, however, everything depends on a hard finite-dimensional choice: the observables must be expressive, nearly invariant under the dynamics, and, ideally, compatible with composition. Deep Koopman methods learn flexible coordinates, whereas structure-preserving methods enforce operator identities on fixed dictionaries. We combine these ideas by introducing Deep Embedded Multiplicative Dynamic Mode Decomposition (DeepMDMD), a method that learns a latent space and a partition of it, while enforcing the Koopman product rule as an exact algebraic constraint. Training alternates between an exact multiplicative operator update and a differentiable latent-clustering step that promotes Koopman closure. The result is a finite transition map on learned latent cells. Its nonzero spectrum lies on the unit circle, its dictionary is shaped by the dynamics rather than by ambient geometry, and forecasts are made in latent coordinates before being decoded to physical space. Across Hamiltonian, chaotic, and fluid examples, DeepMDMD learns dictionaries that are far more compact and dynamically coherent than those produced by geometric MDMD partitions. It reduces spectral pollution, reveals richer continuous-spectrum structure, and gives stable forecasts under severe noise. In high-dimensional flows, including a 158,624-dimensional cylinder wake and a noisy $Re=20,000$ lid-driven cavity, it preserves coherent structures and long-time spectral statistics where state-space MDMD fails. These results suggest a practical rule for Koopman learning: learn the coordinates, constrain the algebra.
arXiv:2606.04867v1 Announce Type: new
Abstract: As AI companion platforms such as Replika and Character.AI rapidly grow, concerns about unsafe human-AI interactions have intensified. This study introduces AICompanionBench, to our knowledge the first publicly available benchmark dataset of human-AI companion conversations annotated with fine-grained safety risk categories. The dataset contains 2,123 real-world Replika conversations collected from Reddit and annotated through human-AI collaboration across nine categories: sexual behavior, antisocial behavior, physical aggression, verbal aggression, substance abuse, self-harm and suicide, control, manipulation, and no-harm. Using this benchmark, we evaluate 20 state-of-the-art open-source and closed-source LLMs under an LLM-as-judge framework for detecting unsafe interactions. Results show substantial variation in model performance, with stronger models achieving high overall accuracy but still struggling with nuanced categories such as manipulation, as well as benign conversations that are incorrectly identified as harmful. Our findings suggest that while current LLMs can effectively detect explicit harmful content, they remain limited in identifying implicit unsafe interactions. Overall, our work contributes a new benchmark dataset for AI companionship safety research and offers insights into monitoring AI companion systems using LLMs. The dataset is publicly available at: https://github.com/anonymousresearcher2026/AICompanionBench/blob/main/AICompanionBench.xlsx
BreastGPT: A Multimodal Large Language Model for the Full Spectrum of Breast Cancer Clinical Routine
arXiv:2606.04911v1 Announce Type: new
Abstract: Breast cancer remains a leading cause of cancer-related mortality among women. Its clinical management requires multimodal reasoning across a clinical workflow that spans \textit{screening}, \textit{diagnosis} and \textit{treatment planning}, where each stage involves distinct imaging modalities, task objectives, and reasoning patterns. However, constrained by data scarcity and model versatility, existing medical MLLMs are typically evaluated on isolated modalities or narrow task families, limiting their ability to support workflow-level clinical reasoning. In this work, we first introduce \textbf{BreastStage}, a workflow-aligned breast imaging instruction corpus comprising 1.86M instruction-following pairs curated from 17 sub-datasets across 5 imaging modalities and 136 task templates. Its held-out split, \textbf{BreastStage-Bench}, provides a comprehensive benchmark for evaluating multimodal reasoning across the breast cancer care continuum. Building on this corpus, we propose \textbf{BreastGPT}, a unified MLLM equipped with a dual-branch visual encoder and concept-preserving token compression to bridge the scale gap between standard radiology and gigapixel pathology. On BreastStage-Bench, BreastGPT achieves 75.66\% closed-ended accuracy and 89.92\% open-ended score, outperforming both general-purpose and medical-specific MLLMs across clinical stages and task formats. These results suggest that workflow-aligned data and cross-scale visual modeling are critical for clinically grounded medical MLLMs. All data, code, and model checkpoints are released at https://yangyy-liu.github.io/BreastGPT.io.
arXiv:2606.04261v1 Announce Type: new
Abstract: Curating training data is among the most consequential yet labor-intensive parts of modern AI development: practitioners iteratively propose, implement, evaluate, and revise data policies against noisy benchmark feedback. We ask whether generalist coding agents can automate this data-curation loop. We introduce *Curation-Bench*, an agent-centric benchmark that fixes the model, training recipe, and evaluation suite while giving agents command-line access to inspect data, implement policies, submit them to a fixed training/evaluation pipeline, and revise. In a vision-language instruction-tuning instantiation, out-of-the-box agents reach strong published data-selection baselines within ten iterations. However, trajectory analysis reveals a persistent *execution-research gap*: agents mainly tune local policy variants rather than explore new policy families, even when given strategy guides and paper references. Scaffolds requiring each iteration to cite, instantiate, and adapt a prior method shift agents toward method-guided exploration. The scaffolded agent autonomously composes -- without human design input -- a data-selection policy that outperforms strong published baselines at one-tenth their data budget. Overall, current agents can run the curation loop, but reliable data research requires scaffolded method adaptation, not open-ended prompting alone. Code and benchmark are open-sourced.
arXiv:2606.04138v1 Announce Type: cross
Abstract: The coordinated interactions between organisms enhance collective functionality, a feature that artificial systems such as enzymatic nanomotors seek to replicate. A key objective, yet still a major challenge, is to achieve chemical communication among nanomotors. Progress has been limited by the difficulties in verifying effective signaling processes, including chemical signal propagation and the response of receiving nanomotors. Here, we address this challenge using an enzymatic nanomotor system that demonstrates communication between two populations through generically non-reciprocal phoretic response. A primary swarm of glucose-responsive nanomotors migrates toward a glucose gradient while producing H2O2 as a diffusible communication signal. This self-generated chemical gradient then acts as a chemoattractant for a secondary swarm of catalase-powered nanomotors. Through carefully designed experiments, we visualize the propagating H2O2 gradient and quantify the spatiotemporal response of the receiver nanomotors to the chemical front. Combined experimental and theoretical analysis has revealed that the synergy between different combinations of chemo-attractive and chemo-repulsive mobilities and catalytic rates of consumption and production of substrates and products gives rise to a wealth of different collective responses in the system. This work represents a step toward programmable synthetic systems at the collective level, broadening the functionality of chemical nanomotors and opening opportunities for future hybrid living-synthetic systems.
arXiv:2606.04174v1 Announce Type: cross
Abstract: Computed Axial Lithography is a volumetric additive manufacturing method that selectively cures photosensitive resin through the 3D superposition of patterns of light, offering advantages over layer-based processes including rapid print times, reduced layer artifacts, and compatibility with high-viscosity materials. However, diffusive effects, primarily those of free-radical quenchers such as oxygen, blur the boundary between cured and uncured regions, limiting resolution and preventing the reproduction of sharp, high-spatial-frequency features. By comparing micro-CT data to computational dose models convolved with kernels across a range of diffusivities, we establish a framework for extracting a single diffusion kernel from any standard uncorrected print to account for all observed deviations from the target. In this work, we correct diffusion-induced blurring by co-optimizing for its effects alongside the inherent blur of the computed tomography reconstruction, demonstrating improved fidelity over previous approaches of pre-compensating the target geometry via deconvolution.
arXiv:2606.04277v1 Announce Type: cross
Abstract: In quantum technologies, continuous-variable systems offer advantages over their discrete counterparts. However, continuous-variable tomography suffers from exponentially growing sample complexity. We propose protocols using quantum mirrors to transfer the complete information of incident photonic states onto a control atomic system. This enables full photonic state characterization through measurements on the control atom alone, realized via kernel functions, direct wavefunction reconstruction, and pointwise Wigner function measurements. Our approach overcomes the limitations of conventional photon counting, statistical inference, and inverse transformation, providing a robust framework for benchmarking and verifying non-Gaussian states in continuous-variable quantum optics.
arXiv:2606.04376v1 Announce Type: cross
Abstract: Real-time multi-camera 3D reconstruction is a key foundation for immersive media, remote interaction and spatial computing. While synchronized camera arrays are widely adopted, achieving geometrically consistent and scalable real-time reconstruction remains challenging. A key challenge is the close linkage among extrinsic calibration, multi-view fusion and global optimization, which causes fluctuating reconstruction results, cumulative errors and poor system expandability. We propose a decoupled framework for calibration and stateless real-time multi-view point cloud fusion (FUSE-Flow), a framework with two collaborative components: geometry-aligned multi-view extrinsic calibration (GMAC) and reliability-guided multi-view point cloud fusion (FUSE). This split design avoids conflicting optimization objectives for targeted improvement. The GMAC module refines camera extrinsics via geometric constraints and multi-view reconstruction transformers, enabling accurate sparse-view calibration without calibration targets, dense images or global bundle adjustment. The FUSE module integrates confidence weighting and adaptive spatial hashing for stateless fusion, ensuring linear time and memory consumption. The two modules mutually reinforce each other: accurate camera poses boost fusion accuracy, and confidence-aware fusion corrects calibration biases. Validated on public datasets and real camera setups, FUSE-Flow outperforms mainstream real-time reconstruction methods in visual effect, dynamic stability and scalability, providing a practical solution for large-scale real-time 3D reconstruction.
arXiv:2606.04868v1 Announce Type: new
Abstract: We report ab initio relativistic calculations on the complex open shell Sn$^{12+}$ highly charged ion, a prototypical plasma ion relevant for extreme ultraviolet (EUV) nanolithography. Previous calculations of EUV emissivity in tin plasmas consistently generate a spectrum in which the region of peak emissivity is blueshifted relative to experiment. By optimising our numerical methods to take full advantage of modern, high-performance CPU architectures, we are able to completely saturate the configuration interaction within the $n=4$ shell. Coupled with a thorough treatment of core-valence correlation we resolve the blueshift, finding a surprisingly large influence of highly excited states on the spectrum that is dominated by multiply excited states.
arXiv:2602.09028v2 Announce Type: replace
Abstract: This paper reformulates Fermat's Last Theorem as an embedding problem of information-geometric structures. We reinterpret the Fermat equation as an $n$-th moment constraint, constructing a statistical manifold $\mathcal{M}_n$ of generalized normal distributions via the Maximum Entropy Principle. By Chentsov's Theorem, the natural metric is the Fisher information metric ($L^2$); however, the global structure is governed by the $L^n$ moment constraint. This reveals a discrepancy between the local quadratic metric and the global $L^n$ structure. We axiomatically define an "Information-Geometric Fermat Solution," postulating that the lattice structure must maintain "dual lattice consistency" under the Legendre transform. We prove the non-existence of such structures for $n \ge 3$. Through the Poisson Summation Formula and Hausdorff-Young Inequality, we demonstrate that the Fourier transform induces an alteration of the function family ($L^n \to L^q$, where $1/n + 1/q = 1$), rendering dual lattice consistency analytically impossible. This identifies a geometric obstruction where integer and energy structures are incompatible within a dually flat space. We conclude by discussing the correspondence between this model and elliptic curves.
arXiv:2602.09075v4 Announce Type: replace
Abstract: In-Context Learning (ICL) in transformers acts as an online associative memory and is believed to underpin their high performance on complex sequence processing tasks. However, in gated linear attention models, this memory has a fixed capacity and is prone to interference, especially for long sequences. We propose Palimpsa, a self-attention model that views ICL as a continual learning problem that must address a stability-plasticity dilemma. Palimpsa uses Bayesian metaplasticity, where the plasticity of each attention state is tied to an importance state grounded by a prior distribution that captures accumulated knowledge. We demonstrate that various gated linear attention models emerge as specific architecture choices and posterior approximations, and that Mamba2 is a special case of Palimpsa where forgetting dominates. This theoretical link enables the transformation of any non-metaplastic model into a metaplastic one, significantly expanding its memory capacity. Our experiments show that Palimpsa consistently outperforms baselines on the Multi-Query Associative Recall (MQAR) benchmark and on Commonsense Reasoning tasks.
arXiv:2606.04019v1 Announce Type: cross
Abstract: Recent studies on sensor-language alignment have shown that two-stage frameworks can improve the semantic modeling ability of wearable-sensor human activity recognition (HAR), where SensorLLM-style methods first perform motion-to-language alignment and then fine-tune the model for downstream tasks. However, our experiments reveal a consistent failure mode when the Stage 2 backbone is compressed to a compact model such as TinyLlama: recognition of dynamic activities remains relatively strong, while the discrimination of low-motion static classes such as standing, sitting, and lying degrades substantially. To address this issue, we propose a gravity-aware hierarchical routing head as a lightweight post-alignment adaptation built on top of an already aligned model, rather than a new large-scale pretraining framework. The method uses the per-channel mean and std from the Chronos tokenizer state to extract statistical cues related to posture and gravity direction, and adaptively combines a static expert and a full expert through soft routing, together with a load-balancing loss for stable training. On the MHealth dataset, this design significantly improves macro-F1 with minimal parameter overhead, and the gains are concentrated mainly on static classes while preserving strong performance on dynamic activities. As a first arXiv disclosure, the current paper reports results on a single dataset only, with the goal of highlighting the core method and laying the groundwork for broader evaluation in future work.
arXiv:2606.04419v1 Announce Type: cross
Abstract: MRI provides excellent soft-tissue contrast without ionizing radiation, but long acquisition times increase patient discomfort while also raising exam costs and limiting scanner throughput. A common approach to reduce scan time is to acquire fewer measurements, which yields an ill-posed linear inverse problem; recovering diagnostic-quality images therefore requires incorporating prior knowledge beyond the measured data. In follow-up exams, the most recent prior scan of a patient can provide a highly informative subject-specific context, but practical use is complicated by temporal changes (including pathology progression), misalignment between scans, and protocol drift across acquisitions. In this work, we introduce L-TGVN, a Longitudinal Trust-Guided Variational Network that leverages prior scans as side information to reconstruct the current scan from heavily undersampled measurements. Crucially, L-TGVN constrains the influence of prior scans to be consistent with the acquired measurements. Unlike many existing longitudinal reconstruction methods, it does not require explicit pre-registration between prior and current scans. It further accommodates differences in acquisition protocols across visits (e.g., changes in sequence parameters). We evaluate L-TGVN against matched-capacity baselines, including prior-guided methods and methods that do not use longitudinal priors, and observe consistent improvements in standard quantitative metrics together with better preservation of fine structures at challenging accelerations. Source code is available at github.com/sodicksonlab/L-TGVN.
arXiv:2606.04429v1 Announce Type: cross
Abstract: A common heuristic used to explain the generalization of first-order gradient methods on non-convex neural networks is that "flat interpolators generalize well" (Hochreiter and Schmidhuber, 1994; Keskar et al., 2017), where flatness can be measured by the trace of the Hessian of the empirical loss. However, Dinh et al. 2017) showed that, using symmetry of the network that can change flatness while keeping the population and empirical losses unchanged, any interpolator can be made sharper or flatter. This result makes the earlier heuristic statement vacuous. In this paper, we show that for learning an unknown multi-index model with $2$-layer non-convex homogeneous neural networks, there is a connection between flatness and generalization, despite the existence of symmetries. This connection pertains to the "flattest" interpolators, i.e., the interpolators that have orderwise minimum flatness among all interpolators. First, we show that there exists a natural class of non-generalizing interpolators whose flatness cannot be made closer to the flattest possible, even using symmetries. Second, we show that for data generated by a sum of single-index models, if the approximation error and label noise are low, any flattest interpolator achieves small population loss, i.e., the flattest interpolators always generalize. This establishes a direct link between flatness and generalization which applies to a large class of activations and realistic data distributions.
arXiv:2606.04444v1 Announce Type: cross
Abstract: Existing datasets cannot support large-scale learning in multi-agent, multi-sensor, or multi-domain autonomy, where diversity and coordination are essential. We present a modular dataset generation pipeline that creates terabyte-scale, ground-truth-labeled data for ground, aerial, and infrastructure-based systems using the AVstack framework and CARLA simulator. Supporting single- and multi-agent configurations with flexible sensor suites, the pipeline enables controllable experimentation across challenging conditions. Representative perception and fusion studies show how generated data can support application-specific training and collaborative autonomy.
arXiv:2606.04576v1 Announce Type: cross
Abstract: Learning Value-at-Risk (VaR) and Expected Shortfall (ES) is important for managing financial risks effectively. Existing approaches with limited parameters are vulnerable to model misspecification in the era of big data. To address this limitation, we propose a large tail risk model, the retrieval-enhanced self-grouping autoencoder (ReSGA), which is designed with millions of parameters to exploit the rich cross-sectional dependence and long-term temporal dynamics of assets using their characteristics. Applied to monthly US equity returns from 1926 to 2023 with 153 firm characteristics, ReSGA outperforms twelve econometric and machine learning competitors in terms of out-of-sample loss and statistical backtesting. In addition, its forecast advantages can translate into significant economic gains from long-short decile portfolios that are constructed by a new size-enhanced left-side momentum strategy. To clarify the role of complexity, we further conduct a systematic scaling analysis and demonstrate that improvements in joint VaR-ES forecasting are primarily driven by data complexity rather than model complexity. Finally, our analyses of group-importance and transfer-learning exhibit the interpretability and cross-market generalizability of ReSGA.
arXiv:2606.04605v1 Announce Type: cross
Abstract: We use lubricated-flow discrete-element-method (LF-DEM) simulations to examine how contact-network organization shapes particle motion in dense shear-thickening suspensions. The primary system studied is a two-dimensional bidisperse monolayer where rigid clusters are identified by the $(3,3)$ pebble game; three-dimensional simulations are shown to have qualitatively similar rotational velocity statistics. Across the stress--solid-fraction state diagram, frictional contact number, $k\ge 3$ percolation, and rigid-cluster fluctuations all strengthen in the same region where translational velocity correlations grow, consistent with rigid clusters translating coherently while the surrounding non-rigid particles accommodate a disproportionate share of the local velocity gradient. Rotational motion provides a complementary view: non-affine angular-velocity distributions broaden, near-contact rotations become increasingly anti-correlated, and rigid and non-rigid particles carry distinct statistics. Connectivity, rigidity, and velocity correlations are related but distinct signatures of the constrained collective motion that accompanies shear-thickening and the approach to shear jamming.
arXiv:2606.04354v1 Announce Type: new
Abstract: Rydberg-atom electrometry is a core technique in the development of highly sensitive quantum electric-field sensors. Its sensitivity based on atom-photon interaction is typically limited by photon shot-noise (PSN) and spectral broadenings. Here, we experimentally demonstrate a near PSN-limited Rydberg electrometry from a 85Rb atomic vapor cell. By engineering atomic coherence through control of residual magnetic fields and laser frequency noise, we achieve the Rydberg electromagnetically induced transparency (EIT) with the narrow linewidth of 1.6 MHz, yielding an enhanced spectral slope for high-sensitivity Rydberg-EIT electrometry. Under optimized superheterodyne detection conditions, we obtain an electric-field sensitivity of 12.5(8) nV cm^-1 Hz^-1/2 at 37 GHz, in close agreement with the calculated PSN limit. These results provide direct experimental evidence of the high-sensitive quantum electrometry and establish a practical route toward quantum-noise-limited Rydberg electrometry.
arXiv:2602.09388v2 Announce Type: replace
Abstract: Multilingual pre-trained language models(mPLMs) offer significant benefits for many low-resource languages. To further expand the range of languages these models can support, many works focus on continued pre-training of these models. However, few works address how to extend mPLMs to low-resource languages that were previously unsupported. To tackle this issue, we expand the model's vocabulary using a target language corpus. We then screen out a subset from the model's original vocabulary, which is biased towards representing the source language(e.g. English), and utilize bilingual dictionaries to initialize the representations of the expanded vocabulary. Subsequently, we continue to pre-train the mPLMs using the target language corpus, based on the representations of these expanded vocabulary. Experimental results show that our proposed method outperforms the baseline, which uses randomly initialized expanded vocabulary for continued pre-training, in POS tagging and NER tasks, achieving improvements by 0.54% and 2.60%, respectively. Furthermore, our method demonstrates high robustness in selecting the training corpora, and the models' performance on the source language does not degrade after continued pre-training.
arXiv:2606.02636v2 Announce Type: replace
Abstract: While sim2real efforts are necessary for effective policy transfer to hardware, there is such a thing as too much of a good thing. We argue that sim2real efforts have led to misaligned incentives with policy learning, resulting in simulator lock in and poor policy exploration due to the unreasonable constraints imposed by the real world. We offer a diagnosis and explanation of the current status of the problem, and propose a potential solution via a sim2sim2real paradigm that leverages the robot's kinematics as the sole design constraint.
arXiv:2602.09464v2 Announce Type: replace
Abstract: Vericoding refers to the generation of formally verified code from rigorous specifications. Recent AI models show promise in vericoding, but a unified methodology for cross-paradigm evaluation is lacking. Existing benchmarks test only individual languages/tools (e.g., Dafny, Verus, and Lean) and each covers very different tasks, so the performance numbers are not directly comparable. We address this gap with AlgoVeri, a benchmark that evaluates vericoding of $77$ classical algorithms in Dafny, Verus, and Lean. By enforcing identical functional contracts, AlgoVeri reveals critical capability gaps in verification systems. While frontier models achieve tractable success in Dafny ($40.3$% for Gemini-3 Flash), where high-level abstractions and SMT automation simplify the workflow, performance collapses under the systems-level memory constraints of Verus ($24.7$%) and the explicit proof construction required by Lean (7.8%). Beyond aggregate metrics, we uncover a sharp divergence in test-time compute dynamics: Gemini-3 effectively utilizes iterative repair to boost performance (e.g., tripling pass rates in Dafny), whereas GPT-OSS saturates early. Finally, our error analysis shows that language design affects the refinement trajectory: while Dafny allows models to focus on logical correctness, Verus and Lean trap models in persistent syntactic and semantic barriers. All data and evaluation code can be found at https://github.com/haoyuzhao123/algoveri.
arXiv:2606.04807v1 Announce Type: new
Abstract: Mitigating social bias in Large Language Models (LLMs) presents a distinct alignment challenge: unlike verifiable tasks, bias lacks a single ground truth, creating a high-variance, subjective reward landscape. Previous preference-based fine-tuning methods have major trade-offs: Direct Preference Optimization (DPO) is limited by the lack of exploration inherent in offline training, while Proximal Policy Optimization (PPO) can lead to training instability due to potentially unreliable critic estimates. In this paper, we propose BiasGRPO, a framework using Group Relative Policy Optimization (GRPO) to stabilize alignment by normalizing rewards across a group of sampled completions. By substituting the value function with a group-relative baseline, our approach reduces instability while maintaining the exploration benefits of online training. We find that BiasGRPO outperforms DPO and PPO across multiple benchmarks, indicating its effectiveness. To adapt GRPO, we synthetically extend a dataset spanning multiple domains and contexts. We also create and release a custom bias reward model that effectively guides generation while being highly compute-efficient and avoiding knowledge degradation, providing a valuable resource that can be seamlessly integrated into multi-objective RLHF pipelines.
arXiv:2606.04872v1 Announce Type: new
Abstract: Cooperative localization (CL) is fundamental in emerging multi-agent systems, where agents fuse local sensing data with exchanged information to estimate their own states. At a large scale, however, tracking cross-correlations becomes infeasible, preventing the use of optimal filters. Ignoring or underestimating these correlations leads to overconfident, and thus inconsistent, estimates. Existing CL algorithms achieve good performance and consistency typically at the expense of communication, computation, or memory that scales with the network size. This is incompatible with ultra large-scale systems (ULSS) - for example, satellite mega-constellations - where per-agent resources are limited and must remain independent of the number of agents. This reveals a critical gap: no existing CL method is simultaneously well-performing, consistent, and ULSS-scalable. This paper introduces a new CL framework that addresses this gap using the recently proposed overlapping covariance intersection methodology, which enables agents to exploit limited structural information about cross-correlations without compromising consistency. The resulting CL algorithm leads to optimal conservative covariance propagation using only locally available information. The method is fully distributed, scalable to an ultra large scale, and provably recursively consistent. Simulations demonstrate substantial performance improvement over state-of-the-art consistent CL approaches while preserving scalability.