Majid Farhadloo, Arun Sharma, Mingzhou Yang
et al.
Traditional foundation models are pre-trained on broad datasets to reduce the training resources (e.g., time, energy, labeled samples) needed for fine-tuning a wide range of downstream tasks. However, traditional foundation models struggle with out-of-distribution prediction and can produce outputs that are unrealistic and physically infeasible. We propose the notation of physics-guided foundation models (PGFM), that is, foundation models integrated with broad or general domain (e.g., scientific) physical knowledge applicable to a wide range of downstream tasks.
Understanding how 5G networks correct errors is no trivial matter. At the heart of the process lie two sophisticated families of codes: LDPC and polar codes. This paper opens the black box, not only by explaining how these codes are designed, but also by showing how they are encoded and decoded in practice. To map where research currently stands, we present a detailed survey of the literature supplemented with insights that are often buried deep within technical standards. These foundations are not just historical footnotes: they are strong candidates for powering error correction in 6G and beyond. In bringing clarity to these building blocks, we aim to help engineers and researchers navigate what is both a complex and increasingly vital part of wireless communication.
Nowadays, sophisticated domains are emerging which require appropriate formalisms to be specified accurately in order to reason about them. One such domain is constituted of smart contracts that have emerged in cyber physical systems as a way of enforcing formal agreements between components of these systems. Smart contracts self-execute to run and share business processes through blockchain, in decentralized systems, with many different participants. Legal contracts are in many cases complex documents, with a number of exceptions, and many subcontracts. The implementation of smart contracts based on legal contracts is a long and laborious task, that needs to include all actions, procedures, and the effects of actions related to the execution of the contract. An ongoing open problem in this area is to formally account for smart contracts using a uniform and somewhat universal formalism. This thesis proposes logical foundations to smart contracts using the Situation Calculus, a logic for reasoning about actions. Situation Calculus is one of the prominent logic-based artificial intelligence approaches that provides enough logical mechanism to specify and implement dynamic and complex systems such as contracts. Situation Calculus is suitable to show how worlds dynamically change. Smart contracts are going to be implement with Golog (written en Prolog), a Situation Calculus-based programming language for modeling complex and dynamic behaviors.
Dana A. Lapides, Gillian Grindstaff, Mary H. Nichols
AbstractFor thousands of years, humans have altered the movement of water through construction of earthworks. These earthworks remain in landscapes, where they continue to alter hydrology, even where structures have long since been abandoned. Management of lands containing earthworks requires an understanding of how the earthworks impact hydrology and knowledge of where the structures are located in the landscape. Various methods for detecting topographic features exist in the literature, including a set of rule and threshold‐based techniques and machine learning methods. These tools are either labor‐intensive or require special pre‐processing or a priori assumptions about structures that limit generalizability. Here, we test a topological analysis tool called “persistence” to determine if it is useful for earthwork detection in rangelands. We found that persistence can be used to detect earthworks with 83% precision and 64% accuracy. Breached berms and berms with significant upslope sedimentation are most likely not to be detected using persistence. These results indicate that persistence can be useful for terrain analysis, and it has the potential to substantially reduce manual effort in feature detection by identifying regions where berms may be found.
Bohmian mechanics supplements the quantum wavefunction with deterministic particle trajectories, offering an alternate, dynamical language for quantum theory. However, the Bohmian wavefunction evolves independently of these trajectories, and is thus unaffected by the observable properties of the system. While this property is widely assumed necessary to ensure agreement with quantum mechanics, much work has recently been dedicated to understanding classical pilot-wave systems, which feature a two-way coupling between particle and wave. These systems—including the “walking droplet” system of Couder and Fort (Couder and Fort (2006) Phys. Rev. Lett. 97:154101) and its various abstractions (Dagan and Bush (2020) CR Mecanique 348:555–571; Durey and Bush (2020) Front. Phys. 8:300; (2021) Chaos 31:033136; Darrow and Bush (2024) Symmetry 16:149)—allow us to investigate the limits of classical systems and offer a touchstone between quantum and classical dynamics. In this work, we present a general result that bridges Bohmian mechanics with this classical pilot-wave theory. Namely, Darrow and Bush ((2024) Symmetry 16:149) recently introduced a Lagrangian pilot-wave framework to study quantum-like behaviours in classical systems; with a particular choice of particle-wave coupling, they recover key dynamics hypothesised in de Broglie’s early double-solution theory (de Broglie (1970) Foundations Phys. 1:5–15). We here show that, with a different choice of coupling, their de Broglie-like system reduces exactly to single-particle Bohmian mechanics in the non-relativistic limit. Our result clarifies that, while multi-particle entanglement is impossible to replicate in general with local, classical theories, no such restriction exists for single-particle quantum mechanics. Moreover, connecting with the previous work of Darrow and Bush, our work demonstrates that de Broglie’s and Bohm’s theories can be connected naturally within a single Lagrangian framework. Finally, we present an application of the present work in developing a single-particle analogue for position measurement in a de Broglie-like setting.
Elena Andreeva, Rishiraj Bhattacharyya, Arnab Roy
et al.
ZK-SNARKs, a fundamental component of privacyoriented payment systems, identity protocols, or anonymous voting systems, are advanced cryptographic protocols for verifiable computation: modern SNARKs allow to encode the invariants of a program, expressed as an arithmetic circuit, in an appropriate constraint language from which short, zero-knowledge proofs for correct computations can be constructed. One of the most important computations that is run through SNARK systems is the verification of Merkle tree (MT) opening proofs, which relies on the evaluation of a fixed-input-length (FIL) cryptographic compression function over binary MTs. As classical, bit-oriented hash functions like SHA-2 are not compactly representable in SNARK frameworks, Arithmetization-Oriented (AO) cryptographic designs have emerged as an alternative, efficient solution. Today, the majority of AO compression functions are built from permutation-based hashing modes, such as Sponge. While this approach allows cost savings, compared to blockcipher-based modes, as it does not require key-scheduling, AO blockcipher schedulers are often cheap to compute. Furthermore, classical bitoriented cryptography has long studied how to construct provably secure compression functions from blockciphers, following the Preneel-Govaerts-Vandewalle (PGV) framework. The potential efficiency gains together with the strong provable security foundations in the classic setting, motivate the study of AO blockcipher-based compression functions. In this work, we propose AO PGV-LC and PGV-ELC, two AO blockcipher-based FIL compression modes inspired by and extending the classical PGV approach, offering flexible input and output sizes and coming with provable security guarantees in the AO setting. We prove the collision and preimage resistance in the ideal cipher model, and give bounds for collision and opening resistance over MTs of arbitrary arity. We compare experimentally the AO PGV-ELC mode over the HADES blockcipher with its popular and widely adopted Sponge instantiation, POSEIDON, and its improved variant POSEIDON2. Our resulting constructions are up to 3× faster than POSEIDONAND 2× faster than POSEIDON2 in native x86 execution, and up to 50% faster in the Groth16 SNARK framework. Finally, we study the benefits of using MTs of arity wider than two, proposing a new strategy to obtain a compact R1CS constraint system in such case. In fact, by combining an efficient parametrization of the HADES blockcipher over the PGV-ELC mode, together with an optimal choice of the MT arity, we measured an improvement of up to 9× in native MT construction time, and up to 2.5× in proof generation time, compared to POSEIDON over binary MTs.
The main contribution of the present paper is the introduction of a simple yet expressive hybrid-dynamic logic for describing quantum programs. This version of quantum logic can express quantum measurements and unitary evolutions of states in a natural way based on concepts advanced in (hybrid and dynamic) modal logics. We then study Horn clauses in the hybrid-dynamic quantum logic proposed, and develop a series of results that lead to an initial semantics theorem for sets of clauses that are satisfiable. This shows that a significant fragment of hybrid-dynamic quantum logic has good computational properties, and can serve as a basis for defining executable languages for specifying quantum programs. We set the foundations of logic programming in this fragment by proving a variant of Herbrand's theorem, which reduces the semantic entailment of a logic-programming query by a program to the search of a suitable substitution.
Neural scaling laws aim to characterize how out-of-sample error behaves as a function of model and training dataset size. Such scaling laws guide allocation of a computational resources between model and data processing to minimize error. However, existing theoretical support for neural scaling laws lacks rigor and clarity, entangling the roles of information and optimization. In this work, we develop rigorous information-theoretic foundations for neural scaling laws. This allows us to characterize scaling laws for data generated by a two-layer neural network of infinite width. We observe that the optimal relation between data and model size is linear, up to logarithmic factors, corroborating large-scale empirical investigations. Concise yet general results of the kind we establish may bring clarity to this topic and inform future investigations.
<p>ATLID (ATmospheric LIDar) is the lidar flown on the multi-instrument Earth Cloud Aerosol and Radiation Explorer (EarthCARE). EarthCARE is a joint ESA–JAXA mission that was launched in May 2024. ATLID is a three-channel, linearly polarized, high-spectral-resolution lidar (HSRL) system operating at 355 nm. Cloud and aerosol optical properties are key EarthCARE products. This paper provides an overview of the ATLID Level 2a (L2a; i.e., single instrument) retrieval algorithms being developed and implemented in order to derive cloud and aerosol optical properties. The L2a lidar algorithms that retrieve the aerosol and cloud optical property profiles and classify the detected targets are grouped together in the so-called A-PRO (ATLID-profile) processor. The A-PRO processor produces the ATLID L2a aerosol product (A-AER); the extinction, backscatter, and depolarization product (A-EBD); the ATLID L2a target classification product (A-TC); and the ATLID L2a ice microphysical estimation product (A-ICE). This paper provides an overview of the processor and its component algorithms.</p>
<p>This paper describes the technical specifications of the extensions made to the middle-atmospheric lidar facility at the Leibniz Institute of Atmospheric Physics in Kühlungsborn, Germany (54.12° N, 11.77° E). The upgrade complements the existing daylight-capable Rayleigh–Mie–Raman (RMR) temperature lidar with a nighttime-only RMR wind–temperature lidar. The new system comprises an independent lidar with laser, telescopes, and detectors, which is synchronized with and adapted to the (old) temperature lidar. As a result, with the combination of RMR lidars the atmosphere is probed with three (vertical and tilted) beams. This work intends to highlight the recent innovations in the construction of a Doppler–Rayleigh lidar system using the single-edge iodine-cell technique, which allows for the simultaneous measurement of wind, temperature, and aerosols. We will detail supporting subsystems that allow for a high degree of lidar automation and concisely provide key technical information about the system that will support readers in the development of additional RMR wind–temperature lidar systems. We show an example of time-resolved temperature and wind soundings reaching up to <span class="inline-formula">∼</span> 90 km. These data agree well with ECMWF-IFS profiles between 35 and <span class="inline-formula">∼</span> 50 km but show a much larger variability above. In the companion article, we will present the algorithm design and uncertainty budgets associated with the data processing chain.</p>
<p>A new algorithm was developed to infer particle size distribution parameters from the Stratospheric Aerosol and Gas Experiment II (SAGE II) and SAGE III on the International Space Station (SAGE III/ISS) extinction spectra using a lookup table (LUT) approach. Here, the SAGE-based extinction ratios were matched to LUT values, and, using these matches, weighted statistics were calculated to infer the median particle size distribution values and higher-moment parameters as well as quantify the uncertainty in these estimates. This was carried out by solving for both single-mode and bimodal lognormal distributions. The work presented herein falls under two general headings: (1) a theoretical study was carried out to determine the accuracy of this methodology, and (2) the solution algorithm was applied to the SAGE II and SAGE III/ISS records with a brief case study analysis of the 2022 Hunga Tonga eruption. This methodology was demonstrated to be <span class="inline-formula">≈</span> 25 % accurate for mode radius and has a minor dependence on particle composition. While bimodal solutions were obtained from this algorithm, we provide a conclusive demonstration of how and why these estimates are inherently unstable using SAGE III/ISS extinction spectra alone. Finally, we demonstrated how the Hunga Tonga aerosol plume evolved in regard to both size and transport over 18 months after the 2022 Hunga Tonga eruption. The particle size distribution (PSD) estimates, higher-moment parameters, and uncertainties are new products within the SAGE III/ISS Level 2 (L2) products, are currently available for download, and will be merged into the main SAGE III/ISS release products in a subsequent L2 release.</p>
<p>The ultraviolet multi-filter rotating shadow-band radiometer (UV-MFRSR) is a seven-channel radiometer with narrowband filters centered between wavelengths 300 and 368 nm. Four of the middle wavelengths in this device are near those used in the Dobson spectrometer to retrieve ozone column abundance. In this paper measurements from Mauna Loa Observatory (MLO) were used first to calibrate the instrument using the Langley plot method and subsequently to derive column ozone and aerosol optical depths. The ozone derived from the UV-MFRSR was compared to the ozone measured by a Dobson spectrophotometer that operates daily at the MLO, resulting in column values within about 1 DU on average for 43 d in 2018. The aerosol optical depth (AOD) retrievals are more challenging. Generally, the AOD increases with wavelength between 305 and 332 nm, not what is expected given the typical AOD wavelength dependence at visible wavelengths. An example of this behavior is discussed, and research by others is cited that indicates similar behavior at these wavelengths, at least for the low-aerosol-optical-depth conditions encountered at high-altitude sites.</p>
We consider consciousness attributed to systems in space-time which can be purely physical without biological background and focus on the mathematical understanding of the phenomenon. It is shown that the set theory based on sets in the foundations of mathematics, when switched to set theory based on ZFC models, is a very promising mathematical tool in explaining the brain/mind complex and the emergence of consciousness in natural and artificial systems. We formalise consciousness-supporting systems in physical space-time, but this is localised in open domains of spatial regions and the result of this process is a family of different ZFC models. Random forcing, as in set theory, corresponds precisely to the random influence on the system of external stimuli, and the principles of reflection of set theory explain the conscious internal reaction of the system. We also develop the conscious Turing machines which have their external ZFC environment and the dynamics is encoded in the random forcing changing models of ZFC in which Turing machines with oracles are formulated. The construction is applied to cooperating families of conscious agents which, due to the reflection principle, can be reduced to the implementation of certain concurrent games with different levels of self-reflection.
Reasoning, a crucial ability for complex problem-solving, plays a pivotal role in various real-world settings such as negotiation, medical diagnosis, and criminal investigation. It serves as a fundamental methodology in the field of Artificial General Intelligence (AGI). With the ongoing development of foundation models, e.g., Large Language Models (LLMs), there is a growing interest in exploring their abilities in reasoning tasks. In this paper, we introduce seminal foundation models proposed or adaptable for reasoning, highlighting the latest advancements in various reasoning tasks, methods, and benchmarks. We then delve into the potential future directions behind the emergence of reasoning abilities within foundation models. We also discuss the relevance of multimodal learning, autonomous agents, and super alignment in the context of reasoning. By discussing these future research directions, we hope to inspire researchers in their exploration of this field, stimulate further advancements in reasoning with foundation models, and contribute to the development of AGI.
Elisabeth Lex, Dominik Kowald, Paul Seitlinger
et al.
Personalized recommender systems have become indispensable in today’s online world. Most of today’s recommendation algorithms are data-driven and based on behavioral data. While such systems can produce useful recommendations, they are often uninterpretable, black-box models, which do not incorporate the underlying cognitive reasons for user behavior in the algorithms’ design. The aim of this survey is to present a thorough review of the state of the art of recommender systems that leverage psychological constructs and theories to model and predict user behavior and improve the recommendation process. We call such systems psychology-informed recommender systems. The survey identifies three categories of psychology-informed recommender systems: cognition-inspired, personality-aware, and affectaware recommender systems. Moreover, for each category, Elisabeth Lex, Dominik Kowald, Paul Seitlinger, Thi Ngoc Trang Tran, Alexander Felfernig and Markus Schedl (2021), “Psychology-informed Recommender Systems”, Foundations and Trends® in Information Retrieval: Vol. 15, No. 2, pp 134–242. DOI: 10.1561/1500000090. Full text available at: http://dx.doi.org/10.1561/1500000090
People say, "A picture is worth a thousand words". Then how can we get the rich information out of the image? We argue that by using visual clues to bridge large pretrained vision foundation models and language models, we can do so without any extra cross-modal training. Thanks to the strong zero-shot capability of foundation models, we start by constructing a rich semantic representation of the image (e.g., image tags, object attributes / locations, captions) as a structured textual prompt, called visual clues, using a vision foundation model. Based on visual clues, we use large language model to produce a series of comprehensive descriptions for the visual content, which is then verified by the vision model again to select the candidate that aligns best with the image. We evaluate the quality of generated descriptions by quantitative and qualitative measurement. The results demonstrate the effectiveness of such a structured semantic representation.