Generative retrieval is an emerging approach in information retrieval that generates identifiers (IDs) of target data based on a query, providing an efficient alternative to traditional embedding-based retrieval methods. However, existing models are task-specific and fall short of embedding-based retrieval in performance. This paper proposes GENIUS, a universal generative retrieval framework supporting diverse tasks across multiple modalities and domains. At its core, GENIUS introduces modality-decoupled semantic quantization, transforming multimodal data into discrete IDs encoding both modality and semantics. Moreover, to enhance generalization, we propose a query augmentation that interpolates between a query and its target, allowing GENIUS to adapt to varied query forms. Evaluated on the M-BEIR benchmark, it surpasses prior generative methods by a clear margin. Unlike embedding-based retrieval, GENIUS consistently maintains high retrieval speed across database size, with competitive performance across multiple benchmarks. With additional re-ranking, GENIUS often achieves results close to those of embedding-based methods while preserving efficiency.
U.S. dollar stablecoins are increasingly used as payment and settlement instruments beyond cryptocurrency markets. With the enactment of the GENIUS Act in 2025, the United States established the first comprehensive federal framework governing their issuance, backing, and supervision. This paper evaluates the financial, technological, and regulatory risks that may arise as GENIUS-compliant stablecoins scale into mainstream use. We show that maintaining par-value redemption may depend not only on backing-asset quality, but also on the functioning of Treasury and repo markets, the balance-sheet capacity of broker-dealers, and the operational reliability of blockchain-based transaction rails. Even conservatively backed stablecoins can face stress from redemption surges, market-intermediation bottlenecks, or technological disruptions. We argue that durable stability will likely require an integrated approach spanning financial-market infrastructure, prudential regulation, and software governance. While grounded in U.S.\ law, the analysis identifies principles that are relevant for regulators in other jurisdictions developing stablecoin regimes.
Advancing LLM reasoning skills has captivated wide interest. However, current post-training techniques rely heavily on supervisory signals, such as outcome supervision or auxiliary reward models, which face the problem of scalability and high annotation costs. This motivates us to enhance LLM reasoning without the need for external supervision. We introduce a generalizable and purely unsupervised self-training framework, named Genius. Without external auxiliary, Genius requires to seek the optimal response sequence in a stepwise manner and optimize the LLM. To explore the potential steps and exploit the optimal ones, Genius introduces a stepwise foresight re-sampling strategy to sample and estimate the step value by simulating future outcomes. Further, we recognize that the unsupervised setting inevitably induces the intrinsic noise and uncertainty. To provide a robust optimization, we propose an advantage-calibrated optimization (ACO) loss function to mitigate estimation inconsistencies. Combining these techniques together, Genius provides an advanced initial step towards self-improve LLM reasoning with general queries and without supervision, revolutionizing reasoning scaling law
The institutionalization of stablecoins has led to a paradigm shift in reserve management, accelerated by the 2025 Green Energy and National Infrastructure Underpinning Stablecoins (GENIUS) Act. This study investigates the "Climate-Liquidity Nexus," defined as the structural vulnerability arising from the use of environmentally sustainable but secondary-market-thin assets as collateral for high-velocity digital payment instruments. Utilizing a Vector Error Correction Model (VECM) and GARCH(1,1) volatility frameworks on high-frequency data from 2024 to 2026, we demonstrate that the transition toward green reserves introduces significant "Liquidity Hysteresis." My empirical results indicate that while green bonds fulfill ESG regulatory mandates, they compromise the information-insensitivity of the 1.00 USD peg. Following exogenous climate-finance shocks, the recovery half-life of green-backed stablecoins is found to be 5.4 times longer than that of traditional Treasury-backed counterparts. We find that the "Greenium" paid by issuers acts as a volatility multiplier rather than a safety buffer. These findings suggest that the current regulatory trajectory may inadvertently catalyze sys
Predictive atomistic simulations have propelled materials discovery, yet routine setup and debugging still demand computer specialists. This know-how gap limits Integrated Computational Materials Engineering (ICME), where state-of-the-art codes exist but remain cumbersome for non-experts. We address this bottleneck with GENIUS, an AI-agentic workflow that fuses a smart Quantum ESPRESSO knowledge graph with a tiered hierarchy of large language models supervised by a finite-state error-recovery machine. Here we show that GENIUS translates free-form human-generated prompts into validated input files that run to completion on $\approx$80% of 295 diverse benchmarks, where 76% are autonomously repaired, with success decaying exponentially to a 7% baseline. Compared with LLM-only baselines, GENIUS halves inference costs and virtually eliminates hallucinations. The framework democratizes electronic-structure DFT simulations by intelligently automating protocol generation, validation, and repair, opening large-scale screening and accelerating ICME design loops across academia and industry worldwide.
Unified Multimodal Models (UMMs) have shown remarkable progress in visual generation. Yet, existing benchmarks predominantly assess $\textit{Crystallized Intelligence}$, which relies on recalling accumulated knowledge and learned schemas. This focus overlooks $\textit{Generative Fluid Intelligence (GFI)}$: the capacity to induce patterns, reason through constraints, and adapt to novel scenarios on the fly. To rigorously assess this capability, we introduce $\textbf{GENIUS}$ ($\textbf{GEN}$ Fluid $\textbf{I}$ntelligence Eval$\textbf{U}$ation $\textbf{S}$uite). We formalize $\textit{GFI}$ as a synthesis of three primitives. These include $\textit{Inducing Implicit Patterns}$ (e.g., inferring personalized visual preferences), $\textit{Executing Ad-hoc Constraints}$ (e.g., visualizing abstract metaphors), and $\textit{Adapting to Contextual Knowledge}$ (e.g., simulating counter-intuitive physics). Collectively, these primitives challenge models to solve problems grounded entirely in the immediate context. Our systematic evaluation of 12 representative models reveals significant performance deficits in these tasks. Crucially, our diagnostic analysis disentangles these failure modes. It
Generative AI (GenAI) has recently emerged as a groundbreaking force in Software Engineering, capable of generating code, identifying bugs, recommending fixes, and supporting quality assurance. While its use in coding tasks shows considerable promise, applying GenAI across the entire Software Development Life Cycle (SDLC) has not yet been fully explored. Critical uncertainties in areas such as reliability, accountability, security, and data privacy demand deeper investigation and coordinated action. The GENIUS project, comprising over 30 European industrial and academic partners, aims to address these challenges by advancing AI integration across all SDLC phases. It focuses on GenAI's potential, the development of innovative tools, and emerging research challenges, actively shaping the future of software engineering. This vision paper presents a shared perspective on the future of GenAI-driven software engineering, grounded in cross-sector dialogue as well as experiences and findings within the GENIUS consortium. The paper explores four central elements: (1) a structured overview of current challenges in GenAI adoption across the SDLC; (2) a forward-looking vision outlining key tec
We introduce GENIUS: a conditional text generation model using sketches as input, which can fill in the missing contexts for a given sketch (key information consisting of textual spans, phrases, or words, concatenated by mask tokens). GENIUS is pre-trained on a large-scale textual corpus with a novel reconstruction from sketch objective using an extreme and selective masking strategy, enabling it to generate diverse and high-quality texts given sketches. Comparison with other competitive conditional language models (CLMs) reveals the superiority of GENIUS's text generation quality. We further show that GENIUS can be used as a strong and ready-to-use data augmentation tool for various natural language processing (NLP) tasks. Most existing textual data augmentation methods are either too conservative, by making small changes to the original text, or too aggressive, by creating entirely new samples. With GENIUS, we propose GeniusAug, which first extracts the target-aware sketches from the original training set and then generates new samples based on the sketches. Empirical experiments on 6 text classification datasets show that GeniusAug significantly improves the models' performance
This report examines the synergy between Large Language Models (LLMs) and Static Application Security Testing (SAST) to improve vulnerability discovery. Traditional SAST tools, while effective for proactive security, are limited by high false-positive rates and a lack of contextual understanding. Conversely, LLMs excel at code analysis and pattern recognition but can be prone to inconsistencies and hallucinations. By integrating these two technologies, a more intelligent and efficient system is created. This combination moves beyond mere vulnerability detection optimization, transforming security into a deeply integrated, contextual process that provides tangible benefits like improved triage, dynamic bug descriptions, bug validation via exploit generation and enhanced analysis of complex codebases. The result is a more effective security approach that leverages the strengths of both technologies while mitigating their weaknesses. SAST-Genius reduced false positives by about 91 % (225 to 20) compared to Semgrep alone.
The GENIUS proposal is described and some of it's physics potential is outlined. Also in the light of the contradictive results from the DAMA and CDMS experiments the Genius TF, a new experimental setup is proposed. The Genius TF could probe the DAMA evidence region using the WIMP nucleus recoil signal and WIMP annual modulation signature simultaneously. Besides that it can prove the long term feasibility of the detector technique to be implemented into the GENIUS setup and will in this sense be a first step towards the realization of the GENIUS experiment.
The status of dark matter search in Heidelberg is reviewed. After one year of running the HDMS prototype experiment in the Gran Sasso Underground Laboratory, the inner crystal of the detector has been replaced with a HPGe crystal of enriched 73Ge. The results of the operation of the HDMS prototype detector are discussed. In the light of the contradictive results from the CDMS and DAMA experiments the GENIUS-TF, a new experimental setup is proposed. The GENIUS-TF could probe the DAMA evidence region using the WIMP nucleus recoil signal and WIMP annual modulation signature simulataneously. Besides that it can prove some key parameters of the detector technique, to be implemented into the GENIUS setup and will in this sense be a first step towards the realization of the GENIUS experiment.
In this short piece, I delved into the connections of Nobel laureates by applying Network Science methods to and public data collected from Wikipedia. I uncovered the existence of a central "giant component" in the Nobel laureate network, highlighting the core-periphery structure and the disparity in visibility among laureates. I explored the dominance of laureates in the fields of science and humanities, revealing a polarization that contradicts the trend of interdisciplinary research. Furthermore, it the finding sheds light on the underrepresentation of female laureates in certain Nobel Prize categories.
The GENIUS (\underline {Ge}rmanium in Liquid \underline {Ni}trogen \underline {U}nderground \underline {S}etup) project has been proposed in 1997 \cite{KK-BEY97} as first third generation double beta decay project, with a sensitivity aiming down to a level of an effective neutrino mass of $<m>\sim$ 0.01 - 0.001 eV. Such sensitivity has been shown to be indispensable to solve the question of the structure of the neutrino mass matrix which cannot be solved by neutrino oscillation experiments alone \cite{KKPS}. It will allow broad access also to many other topics of physics beyond the Standard Model of particle physics at the multi-TeV scale. For search of cold dark matter GENIUS will cover almost the full range of the parameter space of predictions of SUSY for neutralinos as dark matter \cite{KK-Ram,Bed-KK2}. Finally, GENIUS has the potential to be the first real-time detector for low-energy (pp and $^7{Be}$) solar neutrinos \cite{Bau-KK,KKPropos99}. A GENIUS-Test Facility has just been funded and will come into operation by end of 2001.
Double beta decay is indispensable to solve the question of the neutrino mass matrix together with $ν$ oscillation experiments. The most sensitive experiment since eight years - the Heidelberg-Moscow experiment in Gran-Sasso - already now, with the experimental limit of < m_ν> < 0.26 eV excludes degenerate $ν$ mass scenarios allowing neutrinos as hot dark matter in the universe for the small angle MSW solution of the solar neutrino problem. It probes cosmological models including hot dark matter already now on the level of future satellite experiments MAP and PLANCK. It further probes many topics of beyond Standard Model physics at the TeV scale. Future experiments should give access to the multi-TeV range and complement on many ways the search for new physics at future colliders like LHC and NLC. For neutrino physics GENIUS will allow to test almost all neutrino mass scenarios allowed by the present neutrino oscillation experiments. At the same time GENIUS will cover a wide range of the parameter space of predictions of SUSY for neutralinos as cold dark matter. Further it has the potential to be a real-time detector for low-energy (pp and 7Be) solar neutrinos. A GENIUS Te
In this report we describe the implementation and approach developed during the GENIUS Project. The GENIUS project is about the generation of usable user interfaces. It tries to cope with issues related to automatic generation where, usually end-user complain about the poor quality (in term of usability) of generated UI. To solve this issue GENIUS relies on Model-Driven Engineering principles and several MDE tools. Notably, it consists in a set of metamodels specific to the interaction, a set of model transformation embedding usability criteria and an environment for model execution/ interpretation.
GENIUS is a proposal for a large scale detector of rare events. As a first step of the experiment, a small test version, the GENIUS test facility, will be build up at the Laboratorio Nazionale del Gran Sasso (LNGS). With about 40 kg of natural Ge detectors operated in liquid nitrogen, GENIUS-TF could exclude (or directly confirm) the DAMA annual modulation signature within about two years of measurement.
Mendelian randomization (MR) has become a popular approach to study causal effects by using genetic variants as instrumental variables. We propose a new MR method, GENIUS-MAWII, which simultaneously addresses the two salient phenomena that adversely affect MR analyses: many weak instruments and widespread horizontal pleiotropy. Similar to MR GENIUS (Tchetgen Tchetgen et al., 2021), we achieve identification of the treatment effect by leveraging heteroscedasticity of the exposure. We then derive the class of influence functions of the treatment effect, based on which, we construct a continuous updating estimator and establish its consistency and asymptotic normality under a many weak invalid instruments asymptotic regime by developing novel semiparametric theory. We also provide a measure of weak identification, an overidentification test, and a graphical diagnostic tool. We demonstrate in simulations that GENIUS-MAWII has clear advantages in the presence of directional or correlated horizontal pleiotropy compared to other methods. We apply our method to study the effect of body mass index on systolic blood pressure using UK Biobank.
The GENIUS (Germanium in Liquid Nitrogen Underground Setup) project has been proposed in 1997 [KK-BEY97] as first third generation double beta decay project, with a sensitivity aiming down to a level of an effective neutrino mass of <m_ν> < 0.01eV or less. Such sensitivity is important to fix the structure of the neutrino mass matrix with high accuracy, which cannot be done by neutrino oscillation experiments alone. GENIUS will allow broad access also to many other topics of physics beyond the Standard Model of particle physics at the multi-TeV scale. For search of cold dark matter GENIUS will cover a large part of the parameter space of predictions of SUSY for neutralinos as dark matter [Bed-KK2,Ell,KK-LowNu2]. Finally, GENIUS has the potential to be a real-time detector for low-energy (pp and 7{Be}) solar neutrinos [Bau-KK,KK-LowNu2]. A GENIUS-Test Facility has just been funded and will come into operation by end of 2002.
The new project GENIUS will cover a wide range of the parameter space of predictions of SUSY for neutralinos as cold dark matter. Together with DAMA it will be the only experiment which can probe the seasonal modulation signal. Concerning hot dark matter GENIUS will be able to fix the (effective) neutrino mass with high accuracy. A GENIUS Test Facility has just been funded and will come into operation by end of 2002.
GENIUS is a proposal for a large scale detector of rare events like double beta decay, cold dark matter and low-energy solar neutrinos in real time. The idea of GENIUS is to operate a large amount of ``naked'' Ge detectors in liquid nitrogen, with the aim of reducing the background down to a level of 10^(-3) counts/kg keV y. In this work we investigate the contribution to the background of GENIUS coming from argon (39Ar) and krypton (85Kr) contamination in the liquid nitrogen.