Recent advances in large language models (LLMs) enable agentic systems trained with reinforcement learning (RL) over multi-turn interaction trajectories, but practical deployment is bottlenecked by rapidly growing textual histories that inflate token budgets and memory usage. We introduce AgentOCR, a framework that exploits the superior information density of visual tokens by representing the accumulated observation-action history as a compact rendered image. To make multi-turn rollouts scalable, AgentOCR proposes segment optical caching. By decomposing history into hashable segments and maintaining a visual cache, this mechanism eliminates redundant re-rendering. Beyond fixed rendering, AgentOCR introduces agentic self-compression, where the agent actively emits a compression rate and is trained with compression-aware reward to adaptively balance task success and token efficiency. We conduct extensive experiments on challenging agentic benchmarks, ALFWorld and search-based QA. Remarkably, results demonstrate that AgentOCR preserves over 95\% of text-based agent performance while substantially reducing token consumption (>50\%), yielding consistent token and memory efficiency. Our further analysis validates a 20x rendering speedup from segment optical caching and the effective strategic balancing of self-compression.
Wikipedia's founders could not have dreamed they were creating the most important laboratory for social scientific and computing research in history but that is exactly what happened. Hill and Shaw take account of Wikipedia's enormous effect on academic scholarship
Large language models have advanced web agents, yet current agents lack personalization capabilities. Since users rarely specify every detail of their intent, practical web agents must be able to interpret ambiguous queries by inferring user preferences and contexts. To address this challenge, we present Persona2Web, the first benchmark for evaluating personalized web agents on the real open web, built upon the clarify-to-personalize principle, which requires agents to resolve ambiguity based on user history rather than relying on explicit instructions. Persona2Web consists of: (1) user histories that reveal preferences implicitly over long time spans, (2) ambiguous queries that require agents to infer implicit user preferences, and (3) a reasoning-aware evaluation framework that enables fine-grained assessment of personalization. We conduct extensive experiments across various agent architectures, backbone models, history access schemes, and queries with varying ambiguity levels, revealing key challenges in personalized web agent behavior. For reproducibility, our codes and datasets are publicly available at https://anonymous.4open.science/r/Persona2Web-73E8.
While large language models (LLMs) have proven effective in leveraging textual data for recommendations, their application to multimodal recommendation tasks remains relatively underexplored. Although LLMs can process multimodal information through projection functions that map visual features into their semantic space, recommendation tasks often require representing users' history interactions through lengthy prompts combining text and visual elements, which not only hampers training and inference efficiency but also makes it difficult for the model to accurately capture user preferences from complex and extended prompts, leading to reduced recommendation performance. To address this challenge, we introduce HistLLM, an innovative multimodal recommendation framework that integrates textual and visual features through a User History Encoding Module (UHEM), compressing multimodal user history interactions into a single token representation, effectively facilitating LLMs in processing user preferences. Extensive experiments demonstrate the effectiveness and efficiency of our proposed mechanism.
Weiqin Chen, Xinjie Zhang, Dharmashankar Subramanian
et al.
Transformer models (TMs) have exhibited remarkable in-context reinforcement learning (ICRL) capabilities, allowing them to generalize to and improve in previously unseen environments without re-training or fine-tuning. This is typically accomplished by imitating the complete learning histories of a source RL algorithm over a substantial amount of pretraining environments, which, however, may transfer suboptimal behaviors inherited from the source algorithm/dataset. Therefore, in this work, we address the issue of inheriting suboptimality from the perspective of dataset preprocessing. Motivated by the success of the weighted empirical risk minimization, we propose a simple yet effective approach, learning history filtering (LHF), to enhance ICRL by reweighting and filtering the learning histories based on their improvement and stability characteristics. To the best of our knowledge, LHF is the first approach to avoid source suboptimality by dataset preprocessing, and can be combined with the current state-of-the-art (SOTA) ICRL algorithms. We substantiate the effectiveness of LHF through a series of experiments conducted on the well-known ICRL benchmarks, encompassing both discrete environments and continuous robotic manipulation tasks, with three SOTA ICRL algorithms (AD, DPT, DICP) as the backbones. LHF exhibits robust performance across a variety of suboptimal scenarios, as well as under varying hyperparameters and sampling strategies. Notably, the superior performance of LHF becomes more pronounced in the presence of noisy data, indicating the significance of filtering learning histories.
Abinash Kumar Shaw, Raghunath Ghara, Paz Beniamini
et al.
We propose different estimators to probe the intergalactic medium (IGM) during epoch of reionization (EoR) using the dispersion measure (${\rm DM}$) of the fast radio bursts. We consider three different reionization histories, which we can distinguish with a total of $\lesssim 1000\,{\rm DM}$ measurements during EoR if their redshifts are known. We note that the redshift derivatives of ${\rm DM}$ are also directly sensitive to the reionization history. The major point of this work is to explore the variance in the ${\rm DM}$ measurements and the information encoded in them. We find that the all-sky average $\overline{\rm DM}(z)$ gets biased from the line-of-sight (LoS) fluctuations in the ${\rm DM}$ measurements introduced by the ionization of IGM during EoR. We find that the ratio $σ_{\rm DM}/\overline{\rm DM}$ depends directly on the ionization bubble sizes as well as the reionization history. On the other hand, we also find that angular variance (coined as $\textit{structure function}$) of ${\rm DM}$ encodes the information about the duration of reionization and the typical bubble sizes as well. We establish the usefulness of variances in ${\rm DM}$ using toy models of reionization and later verify it with the realistic reionization simulations.
Lekshmi Thulasidharan, Elena D'Onghia, Robert Benjamin
et al.
The prevailing model of galaxy formation proposes that galaxies like the Milky Way are built through a series of mergers with smaller galaxies over time. However, the exact details of the Milky Way's assembly history remain uncertain. In this study, we show that the Milky Way's merger history is uniquely encoded in the vertical thickness of its stellar disk. By leveraging age estimates from the value-added LAMOST DR8 catalog and the StarHorse ages from SDSS-IV DR12 data, we investigate the relationship between disk thickness and stellar ages in the Milky Way using a sample comprising Red Giants (RG), Red Clump Giants (RCG), and metal-poor stars (MPS). Guided by the IllustrisTNG50 simulations, we show that an increase in the dispersion of the vertical displacement of stars in the disk traces its merger history. This analysis reveals the epoch of a major merger event that assembled the Milky Way approximately 11.13 billion years ago, as indicated by the abrupt increase in disk thickness among stars of that age, likely corresponding to the Gaia-Sausage Enceladus (GSE) event. The data do not exclude an earlier major merger, which may have occurred about 1.3 billion years after the Big Bang. Furthermore, the analysis suggests that the geometric thick disk of the Milky Way was formed around 11.13 billion years ago, followed by a transition period of approximately 2.6 billion years leading to the formation of the geometric thin disk, illustrating the galaxy's structural evolution. Additionally, we identified three more recent events -- 5.20 billion, 2.02 billion, and 0.22 billion years ago -- potentially linked to multiple passages of the Sagittarius dwarf galaxy. Our study not only elucidates the complex mass assembly history of the Milky Way and highlights its past interactions but also introduces a refined method for examining the merger histories of external galaxies.
As the final stage of the multi-stage recommender system (MRS), reranking directly affects users' experience and satisfaction, thus playing a critical role in MRS. Despite the improvement achieved in the existing work, three issues are yet to be solved. First, users' historical behaviors contain rich preference information, such as users' long and short-term interests, but are not fully exploited in reranking. Previous work typically treats items in history equally important, neglecting the dynamic interaction between the history and candidate items. Second, existing reranking models focus on learning interactions at the item level while ignoring the fine-grained feature-level interactions. Lastly, estimating the reranking score on the ordered initial list before reranking may lead to the early scoring problem, thereby yielding suboptimal reranking performance. To address the above issues, we propose a framework named Multi-level Interaction Reranking (MIR). MIR combines low-level cross-item interaction and high-level set-to-list interaction, where we view the candidate items to be reranked as a set and the users' behavior history in chronological order as a list. We design a novel SLAttention structure for modeling the set-to-list interactions with personalized long-short term interests. Moreover, feature-level interactions are incorporated to capture the fine-grained influence among items. We design MIR in such a way that any permutation of the input items would not change the output ranking, and we theoretically prove it. Extensive experiments on three public and proprietary datasets show that MIR significantly outperforms the state-of-the-art models using various ranking and utility metrics.
We consider the model of history-deterministic one-counter nets (OCNs). History-determinism is a property of transition systems that allows for a limited kind of non-determinism which can be resolved 'on-the-fly'. Token games, which have been used to characterise history-determinism over various models, also characterise history-determinism over OCNs. By reducing 1-token games to simulation games, we are able to show that checking for history-determinism of OCNs is decidable. Moreover, we prove that this problem is PSPACE-complete for a unary encoding of transitions, and EXPSPACE-complete for a binary encoding. We then study the language properties of history-deterministic OCNs. We show that the resolvers of non-determinism for history-deterministic OCNs are eventually periodic. As a consequence, for a given history-deterministic OCN, we construct a language equivalent deterministic one-counter automaton. We also show the decidability of comparing languages of history-deterministic OCNs, such as language inclusion and language universality.
Dhiraj Kumar Hazra, Daniela Paoletti, Fabio Finelli
et al.
Cosmic Microwave Background temperature and polarization anisotropies from Planck have estimated a lower value of the optical depth to reionization ($τ$) compared to WMAP. A significant period in the reionization history would then fall within $6<z< 10$, where detection of galaxies with Hubble Frontier Fields program and independent estimation of neutral hydrogen in the inter galactic medium by Lyman-$α$ observations are also available. This overlap allows an analysis of cosmic reionization which utilizes a direct combination of CMB and these astrophysical measurements and potentially breaks degeneracies in parameters describing the physics of reionization. For the first time we reconstruct reionization histories by assuming photo-ionization and recombination rates to be free-form and by allowing underlying cosmological parameters to vary with CMB (temperature and polarization anisotropies and lensing) data from Planck 2018 release and a compilation of astrophysical data. We find an excellent agreement between the low-$\ell$ Planck 2018 HFI polarization likelihood and astrophysical data in determining the integrated optical depth. By combining both data, we report for a minimal reconstruction $τ=0.051^{+0.001+0.002}_{-0.0012-0.002}$ at 68\% and 95\% CL, which, for the errors in the current astrophysical measurements quoted in the literature, is nearly twice better than the projected cosmic variance limited CMB measurements. For the duration of reionization, redshift interval between 10\% and complete ionization, we get $2.9^{+0.12+0.29}_{-0.16-0.26}$ at 68\% and 95\% CL, which improves significantly on the corresponding result obtained by using Planck 2015 data. By a Bayesian analysis of the combined results we do not find evidence beyond monotonic reionization histories, therefore multi-phase reionization scenario is disfavored compared to minimal alternatives.
According to what has become a standard history of quantum mechanics, in 1932 von Neumann persuaded the physics community that hidden variables are impossible as a matter of principle, after which leading proponents of the Copenhagen interpretation put the situation to good use by arguing that the completeness of quantum mechanics was undeniable. This state of affairs lasted, so the story continues, until Bell in 1966 exposed von Neumann's proof as obviously wrong. The realization that von Neumann's proof was fallacious then rehabilitated hidden variables and made serious foundational research possible again. It is often added in recent accounts that von Neumann's error had been spotted almost immediately by Grete Hermann, but that her discovery was of no effect due to the dominant Copenhagen Zeitgeist. We shall attempt to tell a story that is more historically accurate and less ideologically charged. Most importantly, von Neumann never claimed to have shown the impossibility of hidden variables tout court, but argued that hidden-variable theories must possess a structure that deviates fundamentally from that of quantum mechanics. Both Hermann and Bell appear to have missed this point, moreover, both raised unjustified technical objections to the proof. Von Neumann's argument was basically that hidden-variables schemes must violate the "quantum principle" that physical quantities are to be represented by operators in a Hilbert space. As a consequence, hidden-variables schemes, though possible in principle, necessarily exhibit a certain kind of contextuality.
We present deep photometry of the Carina dwarf Spheroidal galaxy in the B,V filters from CTIO/MOSAIC, out to and beyond the tidal radius. The accurately calibrated photometry is combined with spectroscopic metallicity distributions of Red Giant Branch stars to determine the detailed star formation and chemical evolution history. The star formation history confirms the episodic formation history of Carina and quantifies the duration and strength of each episode in great detail, as a function radius from the centre. Two main episodes of star formation occurred at old (>8 Gyr) and intermediate (2-8 Gyr) ages, both enriching stars starting from low metallicities ([Fe/H]<-2 dex). By dividing the SFH into two components, we determine that 60pm9 percent of the total number of stars formed within the intermediate age episode. Furthermore, within the tidal radius (0.48 degrees or 888 pc) a total mass in stars of 1.07pm0.08 x10^6 M_sun was formed, giving Carina a stellar mass-to-light ratio of 1.8pm0.8. Combining the detailed star formation history with spectroscopic observations of RGB stars, we are able to determine the detailed age-metallicity relation of each episode and the timescale of alpha-element evolution of Carina from individual stars. The oldest episode displays a tight age-metallicity relation over 6 Gyr with steadily declining alpha-element abundances and a possible alpha-element knee at [Fe/H]~ -2.5 dex. The intermediate age sequence displays a more complex age-metallicity relation starting from low metallicity and a sequence in alpha-element abundances with a slope much steeper than observed in the old episode, starting from [Fe/H]=-1.8 dex and [Mg/Fe]~0.4 dex and declining to Mg-poor values ([Mg/Fe]<-0.5 dex). This indicates clearly that both episodes of star formation formed from gas with different abundance patterns, inconsistent with simple evolution in an isolated system.
The 60th birthday of Johann Rafelski was celebrated during the Strangeness in Quark Matter 2011 in Krakow. Johann was born in Krakow and he initiated the series of the SQM conferences. This report, which briefly presents my personal view on a history of multi-particle production in high energy collisions, is dedicated to Johann.
We carefully analyze how the abundance of Nitrogen over Oxygen evolves when dependent on metallicity stellar yields with a primary component of N proceeding from AGBs stars are used. We show the results obtained with a chemical evolution models grid, calculated with variable star formation efficiencies, which produce different star formation histories. Finally we see how the N/O abundance is related on the evolutionary history.
Star formation happens in a clustered way which is why the star cluster population of a particular galaxy is closely related to the star formation history of this galaxy. From the probabilistic nature of a mass function follows that the mass of the most-massive cluster of a complete population, M_max, has a distribution with the total mass of the population as a parameter. The total mass of the population is connected to the star formation rate (SFR) by the length of a formation epoch. Since due to evolutionary effects only massive star clusters are observable up to high ages it is convenient to use this M_max(SFR) relation for the reconstruction of a star formation history. The age-distribution of the most-massive clusters can therefore be used to constrain the star formation history of a galaxy. The method, including an assessment of the inherent uncertainties, is introduced with this contribution, while following papers will apply this method to a number of galaxies.
Richard Bosworth, The Italian Dictatorship. Problems and Perspectives in the Interpretation of Mussolini and Fascism (London: Arnold, 1998), 269 pp., pb., £14.99, ISBN 0-340-67727-9. Ottar Dahl, Syndicalism, Fascism and Post-Fascism in Italy, 1900–1950 (Oslo: Solum Forlag, 1999), 180 pp., hb., 280 NOK, ISBN 82-560-1187-4. Alexander De Grand, Italian Fascism. Its Origins and Development, 3rd edn (Lincoln, NA, and London: University of Nebraska Press, 2000), 191 pp., pb., £9.95, ISBN 0-8032-6622-7. Dahlia Elazar, The Making of Fascism. Class, State, and Counter-Revolution, Italy 1919–1922 (Westport, CT, and London: Praeger, 2001), 172 pp., hb., $46.50, ISBN 0-275-95864-7. Aristotle Kallis, Fascist Ideology. Territory and Expansionism in Italy and Germany, 1922–1945 (London: Routledge, 2000), 286 pp., pb., £15.99, ISBN 0-415-21612-5. Mark Neocleous, Fascism (Buckingham: Open University Press, 1997), 120 pp., pb., £9.99, ISBN 0-335-19487-7. John Pollard, The Fascist Experience in Italy (London: Routledge 1998), 158 pp., hb., £40.00, ISBN 0-415-11631-7. Jeffrey Schnapp, ed., A Primer of Italian Fascism (Lincoln, NA, and London: University of Nebraska Press, 2000), 325 pp., pb., £16.95, ISBN 0-8032-9268-6.
I discuss the formation of alpha-enhanced metal-rich stellar populations in the nuclei of luminous ellipticals. Based on hierarchical clustering, different galaxy formation scenarios, which imply different star formation histories, are considered. In contrast to the fast clumpy collapse mode, the late merger of two spiral galaxies fails to reproduce significantly $α$-enhanced abundance ratios, unless the IMF is flattened. Following the star formation history predicted by semi-analytic models of hierarchical clustering for the average elliptical, solar abundance ratios are obtained with Salpeter IMF. According to the models, bright ellipticals in the field are expected to have significantly lower Mg/Fe ratios than their counterparts in a cluster.