Search arXiv⌕ Search

arXiv subjects

Cheng Peng

Publications and source records attributed to Cheng Peng.

At least 37 records · Page 2Linked to original sources
arXiv2026

OneLA: Scaling Linear-Attention Decoding to Large Beams in Generative Recommendation

Generative recommendation (GR) relies on large-beam decoding to generate hundreds of candidate items, creating a new scaling challenge for recurrent linear attention. Existing linear attention serving systems either materialize a full recurrent state for every beam or repeatedly replay shared history, incurring substantial memory and traffic overhead. To address this, we present OneLA, a linear-attention decoding framework that exploits the shared prompt and short divergent suffixes of GR workloads. Specifically, OneLA represents all beam states using a single shared prompt-derived state and compact, append-only records of their divergent transitions. Using this representation, OneLA computes only the state information required at each decoding step, without reconstructing a full recurrent state for every beam. Furthermore, OneLA uses a lightweight ancestry index to track the transition records that make up each beam's history, allowing beams to be updated without moving or copying existing records. A fused GPU kernel further reuses the shared state across beams. Our analysis shows that OneLA achieves 1.54-2.46x end-to-end decode speedups while substantially reducing recurrent-state memory use and data movement.

cs.AI↗
arXiv2026

A Rubric-Guided Large Language Model Solution for Opioid Use Disorder Computable Phenotyping

Opioid use disorder (OUD) remains a public health crisis in the United States, yet it is difficult to identify from electronic health records (EHRs) because missing diagnosis codes and supporting evidence are buried in clinical narratives. Accurate OUD identification is critical to support interventions and improve health outcomes. This study developed a rubric-guided large language model (LLM) that incorporated Optimization by PROmpting (OPRO) for OUD computable phenotyping (CP). The framework used an 18-item, expert-identified rubric to instruct LLMs to automatically extract critical text with supporting evidence to determine OUD flags. Two UF Health physicians (GMR and WMG) chart-reviewed 253 patients, including 68 OUD-positive cases. Our LLM-based computable phenotype (CP) achieved the best F1 score of 0.774 and an AUROC of 0.934, outperforming the machine learning-based CP using EHR and natural language processing-extracted variables, and zero-shot LLMs by relative F1 improvements of 12.8% and 44.4%, respectively. The proposed LLM-based CP could link LLM-extracted evidence to OUD phenotyping for better explainability.

cs.CL↗
arXiv2026

World-in-World: World Models in a Closed-Loop World

Generative world models (WMs) can now simulate worlds with striking visual realism, which naturally raises the question of whether they can endow embodied agents with predictive perception for decision making. Progress on this question has been limited by fragmented evaluation: most existing benchmarks adopt open-loop protocols that emphasize visual quality in isolation, leaving the core issue of embodied utility unresolved, i.e., do WMs actually help agents succeed at embodied tasks? To address this gap, we introduce World-in-World, the first open platform that benchmarks WMs in a closed-loop world that mirrors real agent-environment interactions. World-in-World provides a unified online planning strategy and a standardized action API, enabling heterogeneous WMs for decision making. We curate four closed-loop environments that rigorously evaluate diverse WMs, prioritize task success as the primary metric, and move beyond the common focus on visual quality; we also present the first data scaling law for world models in embodied settings. Our study uncovers three surprises: (1) visual quality alone does not guarantee task success, controllability matters more; (2) scaling post-training with action-observation data is more effective than upgrading the pretrained video generators; and (3) allocating more inference-time compute allows WMs to substantially improve closed-loop performance.

cs.CV↗
arXiv2026

An Agentic Generative Large Language Model for Treatment Planning of Colorectal Cancer

Treatment planning in precision oncology requires synthesizing heterogeneous patient information with rapidly evolving clinical guidelines to ensure guideline-concordant care. While large language models (LLMs) show promise in many diagnostic tasks, their adoption for high-stakes treatment planning is hindered by complex reasoning, adherence to timely clinical guidelines, and safety concerns. In this study, we present GatorOnco, an agentic LLM for colorectal cancer (CRC) treatment planning. GatorOnco is developed using a total of 282 billion tokens of biomedical text, including healthcare system-scale clinical text comprising 166 billion tokens from UF Health. We implemented a domain-adaptation method that integrates pre-training, model merging, a two-stage post-training approach, and agent-based reinforcement learning. An agentic retrieval-augmented generation (RAG) approach dynamically integrates time-sensitive clinical guidelines into the reasoning process. In a blind, randomized clinical evaluation conducted by five UF Health oncologists, GatorOnco significantly outperformed open-source LLMs (P < 0.01) and achieved expert-level performance comparable to UF Health oncologists. Compared with expert oncologists, GatorOnco received significantly higher ratings for readability (4.46 vs. 4.19, P < 0.01) and completeness (3.91 vs. 3.52, P < 0.01), while showing statistically comparable performance in correctness (4.09 vs. 4.11, P = 0.921), currency (4.04 vs. 3.98, P = 0.478), and safety (4.22 vs. 4.22, P = 0.999). These findings demonstrate that integrating agentic reasoning with large-scale domain adaptation can help bridge the gap for generative AI in high-stakes cancer treatment planning.

cs.CL↗
arXiv2026

Infinity-Parser2 Technical Report

We present Infinity-Parser2, a large multimodal model that couples a controllable data-synthesis pipeline with multi-task reinforcement learning for end-to-end document parsing, addressing the persistent scarcity of faithfully annotated parsing corpora. Our contributions are threefold. First, we build a scalable synthesis engine, pairing a controllable rendering framework with an iterative refinement loop, and use it to construct and open-source Infinity-Doc2-5M: a 5-million-sample bilingual (Chinese/English) corpus spanning diverse document types, annotated with element bounding boxes, canonical content forms (Markdown, HTML, LaTeX, SMILES, structured charts), and full-page reading order. Second, we introduce a verifiable, multi-task reward system that enables Joint Reinforcement Learning across eight co-trained objectives (document parsing, layout analysis, table parsing, math formula parsing, chart parsing, chemical formula parsing, document VQA, and general multimodal understanding), unifying perception, structure, and reasoning in a single optimization signal. Third, we release two variants under a shared architecture: Infinity-Parser2-Flash, optimized for low-latency inference with a 3.68x throughput gain over Infinity-Parser-7B, and Infinity-Parser2-Pro, engineered for precision-critical settings. Infinity-Parser2-Pro reaches state-of-the-art 87.6% on olmOCR-Bench and 74.3% on ParseBench, surpassing DeepSeek-OCR-2, PaddleOCR-VL-1.5, and MinerU2.5, with strong generalization to charts, chemical formulas, and document VQA.

cs.AI↗
arXiv2026

Anomalous Compressibility and Electronic Robustness of Metallic Delafossite PdCoO$_2$ under Pressure

The layered delafossite PdCoO$_2$ is an exceptional oxide metal whose ultrahigh conductivity arises from a Pd-derived nearly-free-electron band. Here, high-pressure single-crystal X-ray diffraction combined with first-principles calculations is used to investigate its structural, bonding, and electronic evolution up to 10 GPa. PdCoO$_2$ retains the rhombohedral R-3m structure throughout the investigated pressure range, with no structural phase transition. The lattice exhibits an unusual anisotropic compression, with the in-plane a-axis contracting more strongly than the stacking c-axis, opposite to the behavior of most layered materials. Despite this anomalous compressibility, only minor changes are observed in the local Pd-O and Co-O coordination environments, indicating a remarkably rigid bonding framework. Crystal orbital Hamilton population analysis reveals only subtle strengthening of the existing bonding interactions, without the emergence of destabilizing Pd-related antibonding states. Consistent with these findings, the Pd-derived metallic band and quasi-two-dimensional Fermi surface remain essentially unchanged under compression. Boltzmann transport calculations further show that the in-plane conductivity is nearly pressure-independent, whereas the out-of-plane conductivity decreases modestly, resulting in a slight increase in transport anisotropy. These results demonstrate that the anomalous compressibility of PdCoO$_2$ originates from its robust chemical bonding network, which preserves both the crystal structure and the highly conductive Pd-derived metallic state under pressure.

cond-mat.mtrl-sci↗
arXiv2026

Pressure-Stabilized MnSb$_2$ with Complex Incommensurate Magnetic Order

Marcasite-type compounds have been proposed as promising hosts of exotic magnetic quantum states, yet experimental realizations in stoichiometric, disorder-free systems remain limited. Here, we report the high-pressure stabilization and magnetic characterization of MnSb$_2$, a marcasite-type compound that is thermodynamically metastable under ambient pressure. Single crystals were synthesized using a cubic multi-anvil press, and powder and single-crystal X-ray diffraction confirm the orthorhombic $Pnnm$ structure. These crystals are stable at ambient pressure for a long time up to between 450-500 K. Heat-capacity measurements reveal phase transitions at approximately 220 K and 118 K. Neutron diffraction uncovers an unconventional magnetic ground state below 220 K. Magnetic powder neutron diffraction refinements reveal possible multiple magnetic configurations that provide comparably acceptable fits to the experimental data. While most solutions are consistent with a spin-density-wave (SDW) description, helical models systematically yield inferior agreement factors. Across a broad range of models, the Mn ordered moment reaches a maximum value of approximately 2 $μ_B$ and remains predominantly collinear, with minimal canting along the $c$-axis. At 200 K, the magnetic propagation vector is $q$ = (0, 0.3975, 0.3783); upon cooling, the $b$ component increases toward 0.5, reflecting a temperature-dependent evolution of the modulation. The need for modification of the magnetic model between high and low temperatures further highlights the complex and strongly temperature-dependent nature of the magnetic order in this system. These results establish MnSb$_2$ as a pressure-stabilized marcasite magnet with a highly tunable, complex magnetic ground state and a compelling stoichiometric platform for exploring unconventional magnetic behavior, including potential altermagnetism.

cond-mat.str-el↗
arXiv2026

SciFig: Towards Automating Editable Figure Generation for Scientific Papers

High-quality methodology figures are central to scientific communication, yet they remain difficult and time-consuming to create. Such figures must distill a method's components and information flow into a clear, revisable diagram as the paper evolves. Existing methodology diagram automation systems typically face a trade-off between editability and visual quality: TikZ- or SVG-based methods produce editable structured outputs but often lack the richness of human-designed figures, while image-generation models produce polished raster outputs that are difficult to revise. We introduce SciFig, an end-to-end multi-agent framework for generating visually rich and fully editable methodology figures from scientific text. SciFig decomposes figure generation into planning, layout synthesis, component rendering, and iterative refinement, producing XML figures that can be edited in standard diagramming tools and refined through human or VLM feedback. We also introduce SciFig-Bench, a human-verified benchmark of 435 author-drawn methodology figures from 37 arXiv domains and 15 top-tier AI/ML venues, and SciFig-Eval, a four-axis evaluation protocol for measuring figure quality. Across seven single-agent and agentic baselines, SciFig achieves the best performance on all four SciFig-Eval axes and generates editable figures in about 10 minutes on average. Qualitative examples further show that SciFig can generalize beyond methodology figures to teaser diagrams and statistical plots. Dataset and code are available at: https://shramanpramanick.github.io/SciFig/.

cs.AI↗
arXiv2026

Boron-assisted synthesis of compositionally complex amorphous oxides via short-range-order-constrained generative design

Engineering short-range atomic order in amorphous materials offers a promising yet still underexplored route to high-performance solids. Here, we establish a boron-assisted amorphization strategy through ApolloX, a theory-guided, short-range-order-constrained generative framework for identifying low-energy amorphous configurations in compositionally complex multimetal BOx systems. Using FeCoNiMoBOx as a representative model platform, ApolloX predicts an ensemble of candidate amorphous configurations across systematically varied boron contents. Ab initio molecular dynamics simulations based on these configurations show that increasing boron content suppresses atomic diffusion and disfavors crystallization, with the stabilization of BO3-centered local motifs emerging as a key structural feature associated with enhanced amorphization propensity. Guided by these predictions, we synthesize three representative FeCoNiMoBOx compositions with distinct boron contents and use synchrotron-based scattering and electron microscopy to verify compositional fidelity, structural homogeneity, and the targeted amorphous features, thereby experimentally validating the boron-regulated structural evolution predicted by theory. Beyond this representative system, the same strategy is extended to a broader library of multimetal BOx amorphous compositions spanning diverse metal combinations and boron loadings, demonstrating that the boron-assisted route is not limited to a single FeCoNiMoBOx family but is broadly transferable across compositionally complex amorphous oxides. Overall, our results establish boron incorporation as a practical design variable for tuning short-range order and amorphization in multicomponent oxides, and provide a general framework for the theory-guided discovery of compositionally complex amorphous materials with tunable properties.

cond-mat.mtrl-sci↗
arXiv2026

Prompt, Plan, Extract: Zero-Shot Agentic LLMs Workflows for Lung Pathology Extraction from Clinical Narratives

Information extraction from pathology reports is essential for cancer staging, tumor registry population. Yet key data remains embedded in narrative reports, making manual extraction labor-intensive and error-prone. Traditional supervised Natural Language Processing pipelines address this through fully supervised Named Entity Recognition and Relation Extraction, but require expensive manual annotation and suffer cascading failures when upstream entities are missed. In this study, we developed a zero-shot, agentic workflow, and evaluated five open-source generative Large Language Models (LLMs) to populate 13 College of American Pathologists synoptic fields from lung resection pathology reports. We compared them against a state-of-the-art supervised GatorTron NER-RE baseline using a novel, registry-aligned evaluation framework. The baseline achieved Micro-F1of 0.960, while the best zero-shot model (GPT-OSS-20B) achieved Micro-F1 of 0.893 (recall: 0.949), accurately extracting complex relations like Pathologic Stage without task-specific training. These results suggest that open-source, zero-shot agentic LLMs show great potential as a low-cost solution for extracting lung pathology information.

cs.CL↗
arXiv2026

Automatic Summarization of Doctor-Patient Encounter Dialogues Using Large Language Model through Prompt Tuning

Automatic text summarization (ATS) is an emerging technology to assist clinicians in providing continuous and coordinated care. This study presents an approach to summarize doctor-patient dialogues using generative large language models (LLMs). We developed prompt-tuning algorithms to instruct generative LLMs to summarize clinical text. We examined the prompt-tuning strategies, the size of soft prompts, and the few-short learning ability of GatorTronGPT, a generative clinical LLM developed using 277 billion clinical and general English words with up to 20 billion parameters. We compared GatorTronGPT with a previous solution based on fine-tuning of a widely used T5 model, using a clinical benchmark dataset MTS-DIALOG. The experimental results show that the GatorTronGPT- 20B model achieved the best performance on all evaluation metrics. The proposed solution has a low computing cost as the LLM parameters are not updated during prompt-tuning. This study demonstrates the efficiency of generative clinical LLMs for clinical ATS through prompt tuning.

cs.CL↗
arXiv2026

PIGEON: VLM-Driven Object Navigation via Points of Interest Selection

Object navigation in unseen indoor environments requires agents to perform semantic search under partial observability. Vision-language models (VLMs) provide strong semantic-spatial priors for this task, but how to interface them with robot navigation remains challenging: dense VLM inference is expensive, while abstracting environments into symbolic memories often separates high-level reasoning from the raw visual evidence that supports it. We propose we propose PIGEON (Point of Interest Guided Exploration for Object Navigation), a VLM-driven framework that formulates object navigation as raw-observation-grounded sparse decision problem. PIGEON introduces Points of Interest (PoIs) as sparse visual decision units that couple geometrically executable waypoints with raw egocentric observations. Rather than using VLMs as dense controllers or restricting them to frontier ranking, PIGEON enables VLMs to select among task-critical PoIs, including exploration frontiers, suspected target objects, traversable stairs, and floor-level summaries, while low-level planners execute continuous motion between them. This PoI interface further makes high-level navigation decisions verifiable, allowing us to develop an RLVR pipeline that improves local VLMs without manual Chain-of-Thought annotations. Extensive experiments on Habitat ObjectNav benchmarks show that PIGEON achieves state-of-the-art zero-shot performance, scales consistently with foundation model capacity, and transfers to Active Embodied Question Answering with only prompt modifications. Real-world deployments on physical robots further demonstrate its robustness and efficiency.

cs.RO↗
arXiv2026

Site Preferences and "Coloring Problem" in Cu-doped BiMn$_7$O$_{12}$ Quadruple Perovskite

Lightly Cu-doped BiMn$_7$O$_{12}$ (x = 0.05, 0.10, and 0.15) was investigated using high-pressure synthesis, single-crystal X-ray diffraction, pair distribution function (PDF) analysis, STEM, magnetic measurements, and first-principles calculations. All compositions retain an average monoclinic $I$2/$m$ structure, while Cu substitution progressively suppresses the monoclinic distortion and drives the lattice toward a pseudo-cubic metric symmetry. PDF analysis reveals increasing local structural disorder and reduced medium-range coherence with increasing Cu concentration, despite preservation of the overall quadruple-perovskite framework. Single-crystal refinements indicate enhanced electron density at the octahedral Mn B sites, suggesting preferential Cu occupation within the MnO$_6$ network rather than the conventional square-planar sites expected for Cu$^{2+}$. Magnetic measurements reveal two characteristic anomalies near $T_1$ ~ 100-120 K and $T_2$ ~ 50-60 K, together with pronounced magnetic irreversibility, field-dependent hysteresis, and unsaturated magnetization. Increasing Cu concentration progressively suppresses the low-temperature magnetic state and weakens the field-induced moment. First-principles calculations favor Cu occupation at the square-planar sites, contrasting with the experimental refinements and highlighting strong competitions among local bonding, short-range disorder, and metastability in this highly frustrated quadruple perovskite system.

cond-mat.mtrl-sci↗
arXiv2026

Extending the La Solubility Limit in Sr$_3$Ir$_2$O$_7$ through High-Pressure High-Temperature Synthesis

La-doped bilayer iridates provide an important platform for studying the evolution of the spin-orbit-assisted Mott state under electron doping, but the La solubility achieved by conventional ambient-pressure synthesis is limited. Here, we report the synthesis and physical properties of nominally La-doped (Sr$_{1-x}$La$_x$)$_3$Ir$_2$O$_7$ (x = 0.05, 0.10, 0.15, and 0.20) prepared using high-pressure high-temperature techniques. Single-crystal X-ray diffraction refinements, supported by scanning electron microscopy-energy-dispersive X-ray spectroscopy (SEM-EDX), reveal significantly enhanced La incorporation, with nominal x = 0.05 and 0.15 corresponding to actual compositions of approximately (Sr$_{0.89}$La$_{0.11}$)$_3$Ir$_2$O$_7$ and (Sr$_{0.77}$La$_{0.23}$)$_3$Ir$_2$O$_7$, respectively. At nominal x = 0.20, the bilayer phase is no longer stabilized and instead transforms into cubic perovskite Sr$_{1-x}$La$_x$IrO$_3$. (Sr$_{0.89}$La$_{0.11}$)$_3$Ir$_2$O$_7$ exhibits a ferromagnetic-like transition near 186 K accompanied by magnetic hysteresis and subtle lattice anomalies indicative of spin-lattice coupling. Despite its high electron-doping level, the compound remains strongly insulating, consistent with a heavily doped localized magnetic insulating state distinct from both parent Sr$_3$Ir$_2$O$_7$ and ambient-pressure La-doped samples. In contrast, (Sr$_{0.77}$La$_{0.23}$)$_3$Ir$_2$O$_7$ displays metal-like electronic behavior, weakened magnetic order, and enhanced carrier delocalization, although disorder-driven localization persists at low temperatures. These results demonstrate that high-pressure synthesis substantially extends the accessible doping range of bilayer iridates and reveals electronic and magnetic states inaccessible through conventional synthesis routes.

cond-mat.mtrl-sci↗
arXiv2026

Multi-resolution Enhancement for Full Spectrum Neural Representations

Scientific data acquisition continues to outpace storage and analysis capabilities, making voxel-based representations increasingly intractable. Implicit neural representations (INRs) offer a promising solution by encoding signals through coordinate-based neural networks, serving as surrogates of data, with computational and storage requirements scaling with network complexity rather than data dimensionality. However, smaller INRs struggle to faithfully represent the multi-scale structures, high-frequency information, and fine textures that constitute a large proportion of scientific measurements. We propose WIEN-INR, a theoretically-guided hierarchical INR framework that distributes modeling across resolution scales and enables improved representation capacity through a novel enhancement network to recover subtle details. This multi-scale architecture allows smaller networks to retain the full spectrum of information while preserving training efficiency and lowering storage cost. Evaluated on distinct raw experimental measurements across scales and complexities, WIEN-INR represents a practical step toward broader adoption of neural representations in scientific workflows, delivering compact, robust, and high-fidelity representations.

cs.LG↗
arXiv2026

Sequential Bootstrap for Out-of-Bag Error Estimation: A 100-Seed Replication Study and Variance-Structure Analysis

Out-of-Bag (OOB) estimation is the standard internal diagnostic for bootstrap-aggregated tree ensembles. Under the classical multinomial bootstrap, the number of distinct training observations in each replicate, $U_b$, is itself random, but its contribution to OOB-based variability has rarely been isolated empirically. We use Sequential Bootstrap (SB) -- a resampling scheme that holds $U_b$ at a fixed target $k_n = \lfloor 0.632 n\rfloor$ -- as a controlled perturbation of the bootstrap mechanism, and ask whether stabilizing $U_b$ produces any measurable change in OOB-based diagnostics. We reproduce Breiman's five OOB experimental families on twelve synthetic and real datasets, but unlike the three-seed presentation common in this literature, we run 100 independent random seeds with 50 internal replications per seed, enabling formal paired statistical comparison (Wilcoxon signed-rank, paired-$t$, Pitman--Morgan variance test). We report three findings. First, OOB means are essentially insensitive to stabilization of $U_b$: of 57 (experiment, dataset, metric) cells under 100 seeds, only 6 reach $p<0.05$ on the paired mean comparison, and 4 of those 6 point in the opposite direction from what a 3-seed reading would suggest. Second, a narrow but reproducible effect survives at the variance level: SB reduces the cross-seed standard deviation of node-level classification diagnostics on real datasets while slightly increasing it on synthetic ones (permutation $p=0.026$); the Vehicle dataset exhibits a 21% cross-seed sd reduction (Pitman--Morgan $p=0.017$). Third, several directional claims that appear stable across three seeds flip sign under 100-seed replication, illustrating the cost of underpowered replication protocols. We therefore treat SB as a diagnostic tool for probing the distinct-sample-count term in the variance of OOB estimators, not as an alternative to the classical bootstrap.

stat.ME↗
arXiv2026

Compressibility and High-Pressure Structure of CaMg$_2$Bi$_2$ and YbMg$_2$Bi$_2$

Compounds with the formula $AM_2X_2$ in the CaAl$_2$Si$_2$ structure type have garnered increasing interest across various solid-state research domains, such as quantum topological and thermoelectric materials. Prior studies have identified high-pressure phase transitions in several compounds, including Mg$_3$Sb$_2$, Mg$_3$Bi$_2$, CaMn$_2$Bi$_2$, and SrAl$_2$Si$_2$. In this study, we investigate the structural behavior of CaMg$_2$Bi$_2$ and YbMg$_2$Bi$_2$ under varying pressure conditions. We synthesized crystals using the molten metal flux method and examined them through single-crystal synchrotron X-ray diffraction, employing diamond anvil cells to exert pressures up to 20 GPa. Our analysis reveals insights into the anisotropic compressibility of these materials, highlighting the more compressible and flexible octahedral $A$-Bi bonds as the primary contributors to this anisotropy. Moreover, we observed a phase transition in both CaMg$_2$Bi$_2$ and YbMg$_2$Bi$_2$ at pressures above 9.6 GPa and 8.7 GPa, respectively. The newly identified high-pressure phase exhibits a distortion of the original CaAl$_2$Si$_2$ structure with space group $C2/m$. This high-pressure structure is distinct from that of related compounds (e.g., CaMn$_2$Bi$_2$, MgMg$_2$Bi$_2$), the latter exhibiting a square pyramidal coordination for the $M$ site.

cond-mat.mtrl-sci↗
arXiv2026

Prior-Agnostic Robust Forecast Aggregation

Robust forecast aggregation combines the predictions of multiple information sources to perform well in the worst case across all possible information structures. Previous work largely focuses on settings with a known binary state space, where the state is either 0 or 1. We study prior-agnostic robust forecast aggregation in which the aggregator observes only experts' reports, yet is ignorant of both the underlying joint information structure and the full prior, including the underlying state space. Unlike the standard model that fixes the binary state space {0, 1}, we allow the (binary) unknown state values to be arbitrary numbers in [0, 1], so the same reported probability may correspond to very different realized outcome frequencies across environments. Our main contribution is a simple, explicit, closed-form log-odds aggregator that linearly pools forecasts in logit space, together with (nearly-)tight minimax-regret guarantees across three knowledge regimes. We first show that under conditionally independent (CI) signals, robust aggregation with an unknown state space is strictly harder than in the known-state setting by establishing a larger lower bound, and our aggregation rule can achieve a worst-case regret of 0.0255. Along the way, we also characterize tight regret bounds for Blackwell-ordered structures and for general information structures. In the classical setting with known state space {0,1}, our aggregator achieves regret strictly below 0.0226 for CI structures. To the best of our knowledge, this is the first explicit closed-form aggregator that achieves a regret upper bound strictly less than 0.0226. Finally, we extend the model where the aggregator additionally knows each expert's marginal forecast distribution; in this setting, with the CI structures, we show that a generalized log-odds rule achieves regret of 0.0228, complementing with a lower bound of 0.0225.

cs.LG↗