We propose a time-multiplexed photonic network architecture based on coupled ring resonators, capable of accurately emulating specific Hamiltonian dynamics. We show that, in the Suzuki-Trotter limit, the resulting stroboscopic evolution reproduces the characteristic dynamics of the bosonized Hopfield model. Furthermore, by incorporating a nonlinear element within the main resonator loop, we outline a scalable route toward optical simulation of both mean-field and quantum nonlinear dynamics associated with the Tavis-Cummings model. Our results establish time-multiplexed resonator networks as a versatile photonic framework for simulating interacting light-matter Hamiltonians and collective many-body phenomena.

The large-$N$ quantum field theories provide a window into the regime of strongly-coupled physics. Our principal object of study in this thesis is the large-$N$ family of melonic QFTs, which contain the Sachdev-Ye-Kitaev-like models, tensor models, and vector models. We begin with a review of this limit of a large number of degrees of freedom (large-$N$) as an approach to the solution of QFTs. Two toy models are used to clarify this approach: a zero-dimensional field theory and the flow of a generalized free field theory. Both models are solvable, and so we can explicitly demonstrate: using the former, the simplifications at large $N$; using the latter, the tools used to study scale-dependence of physics -- the renormalization group. We develop $\tilde{F}$-extremization, a simple method of solution for an arbitrary large-$N$ melonic QFT in its strongly-coupled limit. The infrared conformal field theories show remarkable simplicity, in that they are entirely solved by the requirement that they have as many degrees of freedom as possible, up to a simple constraint arising from the interaction between the fields. We measure the number of degrees of freedom of the conformal infrared theory via $\tilde{F}$, the universal part of the free energy. We then present the example of the quartic Yukawa model in continuous dimension. This model is considered as a tensor field theory, and solved for its conformal limit; we then illustrate its multiplicity of fixed points and their stability, as well as its operator spectrum, matching the data between the large-$N$ and dimensional expansions. These features reflect general characteristics of melonic conformal field theories: their existence, stability, and spectral characteristics. We conclude with future directions of exploration for the melonic theories.

Modern EDA flows rely heavily on Tcl scripting, yet general LLMs perform poorly in this domain due to extreme data scarcity, domain-specific semantics, and the high reliability required in physical design. We present iScript, a domain-adapted Qwen3-8B model for Innovus Tcl script generation, and iScript-Bench, a comprehensive benchmark covering five task categories and three difficulty levels. To overcome the lack of training data, we introduce a multi-stage data synthesis pipeline that integrates command extraction, static linting, requirement back-inference, and Chain-of-Thought generation, producing a 10K-tuple (requirement, CoT, script) dataset. iScript is trained through a two-stage strategy combining domain-adaptive pretraining and supervised fine-tuning. To evaluate script correctness efficiently, we further propose a two-step verification framework consisting of static syntax verification and LLM-based functional evaluation. On our benchmark, iScript shows higher pass@k scores than currently state-of-the-art LLMs on average. These results demonstrate the effectiveness of domain adaptation and data synthesis for EDA scripting tasks.

A PLASEN (Precision LAser Spectroscopy for Exotic Nuclei) system, consisting of a compact radio-frequency quadrupole cooler-buncher (RFQ-cb) and a collinear resonance ionization spectroscopy setup, has now been fully commissioned with radioactive ion beams at the Beijing Radioactive Ion-beam Facility (BRIF). Using both stable and radioactive Rb ion beams from BRIF, we demonstrated that the large beam energy spread observed at BRIF has been successfully handled by employing the RFQ-cb, enabling the delivery of high-quality bunched radioactive ion beams for collinear resonance ionization spectroscopy experiments. Under these conditions, we performed laser spectroscopy of exotic nuclei, achieving high resolution (about 100 MHz spectral linewidth) and high sensitivity (up to 1:200 efficiency). This fully operational PLASEN system will serve as a state-of-the-art experimental platform at BRIF for research in multiple fields such as nuclear, atomic and molecular physics.

Neutron-induced reactions play an important role in fundamental nuclear physics, nuclear astrophysics, and applications. In the case of reactions on rare isotopes, there are limited options for direct experimental measurements. The Neutron Target Demonstrator project at Los Alamos National Laboratory seeks to test the feasibility of moderating spallation neutrons within a 1~m$^3$ graphite cube to create a standing neutron target for neutron-induced reaction measurements in inverse kinematics. This paper presents the results of experimental neutron flux distribution tests using neutron sources (ranging from 1~keV to 50~MeV) created by accelerators at the University of Notre Dame and Texas A\&M University. Measurements were made with both the full graphite cube as well as a ''half cube'' setup in which half of the graphite cube was removed. The measured distributions agree with simulated distributions in the case of the full cube moderator, although there remain discrepancies in certain cases for the half cube moderator. The results shown here will provide useful information for an upcoming experimental campaign to test the neutron target proof-of-principle.

Multi-Agent System is emerging as the \textit{de facto} standard for complex task orchestration. However, its reliance on autonomous execution and unstructured inter-agent communication introduces severe risks, such as indirect prompt injection, that easily circumvent conventional input guardrails. To address this, we propose \SysName, a framework that shifts the defensive paradigm from static input filtering to execution-aware analysis. By extracting and reconstructing Cross-Agent Semantic Flows, \SysName synthesizes fragmented operational primitives into contiguous behavioral trajectories, enabling a holistic view of system activity. We leverage a Supervisor LLM to scrutinize these trajectories, identifying anomalies across data flow violations, control flow deviations, and intent inconsistencies. Empirical evaluations demonstrate that \SysName effectively detects over ten distinct compound attack vectors, achieving F1-scores of 85.3\% and 66.7\% for node-level and path-level end-to-end attack detection, respectively. The source code is available at https://anonymous.4open.science/r/MAScope-71DC.

The MexNICA Collaboration coordinates the activities of Mexican scientists, engineers, postdoctoral fellows and students in the Multi-Purpose Detector experiment at the Nuclotron-based Ion Collider fAcility of the Joint Institute for Nuclear Research in Dubna, Russia. Established in 2016, the collaboration brings together five Mexican institutions whose contributions span detector development as well phenomenological and theoretical studies, including modeling by means of Monte Carlo simulations. This work summarizes the main achievements of MexNICA, consisting of the development of the miniBeBe trigger detector as well of results of phenomenological investigations of the baryon-rich region in the QCD phase diagram accessible at NICA energies, and theoretical advances based on lattice QCD and effective models.

This is a document discussing the creation and usage of a server system dedicated to retrieving, processing, and storing data generated from the Solar Neutrino and Astro-Particle PhYsics (SNAPPY) CubeSat by the nuSOL (Neutrino Solar Orbiting Laboratory) Project. On a traditional desktop computer with CERN's ROOT and PostgreSQL software installed, and with a file system on two mirrored drives, it is possible to automatically process and organize incoming data, along with keeping a database to record each incoming file along with a command record. In addition to this, an application was created to provide a Graphical User Interface to assist with creating commands to communicate with the CubeSat. With that said, there are still plenty of plans to improve the software, mainly providing an automatic emailing system to notify team members when they are not around the server.

Full-dimensional quantum scattering calculations are reported for ro-vibrational transitions in HD+HD collisions using a highly accurate interaction potential for the H$_2$-H$_2$ system. Several near-resonant ro-vibrational transitions are identified that conserve the overall rotational angular momentum and nearly conserve the internal energy of the collision partners. Key anisotropic terms that drive the rotational transitions and angular momentum partial waves that contribute to low energy resonant features in the energy dependence of the cross sections are identified. The computed results are in agreement with total cross sections reported in previous experimental results, including resonant features in the energy dependence of the cross section. In particular, low-energy cross sections show a strong resonant feature associated with an $l=3$ partial wave in the incident channel. Rate coefficients for several inelastic rotational and ro-vibrational transitions are reported for temperatures ranging from $0.1$ K to $200$ K and they display a maximum between $1$ K-$10$ K reflecting the important contributions from the $l=3$ shape resonance that occurs around 2.5 K.

The proliferation of deepfake imagery poses escalating challenges for practitioners tasked with verifying digital media authenticity. While detection algorithm research is abundant, empirical evaluations of publicly accessible tools that practitioners actually use remain scarce. This paper presents the first cross-paradigm evaluation of six tools, spanning two complementary detection approaches: forensic analysis tools (InVID \& WeVerify, FotoForensics, Forensically) and AI-based classifiers (DecopyAI, FaceOnLive, Bitmind). Both tool categories were evaluated by professional investigators with law enforcement experience using blinded protocols across datasets comprising authentic, tampered, and AI-generated images sourced from DF40, CelebDF, and CASIA-v2. We report three principal findings: forensic tools exhibit high recall but poor specificity, while AI classifiers demonstrate the inverse pattern; human evaluators substantially outperform all automated tools; and human-AI disagreement is asymmetric, with human judgment prevailing in the vast majority of discordant cases. We discuss implications for practitioner workflows and identify critical gaps in current detection capabilities.

We present Threadle, an open-source, high-performance, and memory-efficient network storage and query engine written in C#. Designed for working with full-population networks derived from administrative register data, which represent very large, multilayer, mixed-mode networks with millions of nodes and billions of edges, Threadle addresses a fundamental limitation of existing network libraries: the inability to efficiently handle two-mode (bipartite) data at scale. Threadle's core innovation is a pseudo-projection approach that allows two-mode layers to be queried as if they were projected into one-mode form, without ever materializing the memory-prohibitive projection. We demonstrate that a network with 20 million nodes containing layers equivalent to 8 trillion projected edges can be stored in approximately 20 GB of RAM -- a compression ratio exceeding 2000:1 compared to materialized projection. Additionally, Threadle provides native support for multilayer mixed-mode networks, an integrated node attribute manager, and a CLI frontend with 50+ commands for the construction, processing, file handling, and management of very large heterogeneous networks. Threadle is freely available at https://www.threadle.dev and can either be obtained as precompiled binaries for Win, macOS and Linux, or compiled directly from source. Supplementing Threadle is threadleR, an R frontend that enables advanced sampling- and traversal-based analyses on very large, heterogeneous, multilayer, mixed-mode population-scale networks.

Traditional macro-cell and micro-cell infrastructures suffer from severe inefficiencies, with current macro-cell networks operating at less than 5 percent energy efficiency, leading to nearly 95 percent of RF power wasted in covering vacant areas. The problem becomes particularly acute in high-density scenarios such as the Hajj, where approximately 7,000 temporary diesel-powered towers are deployed each year, consuming 56 million liters of fuel and emitting around 148,000 tons of CO2, yet still experiencing failure rates of nearly 40 percent at peak demand. To overcome these limitations, we propose an AI-driven mesh architecture based on three integrated enablers: (i) proximity-based deployment of low-power nodes within 250 to 300 meters of users, yielding a 38 dB link-budget gain and up to 6000 times efficiency improvement; (ii) spatial frequency reuse, which partitions cells into multiple non-interfering zones and achieves nearly 20 times capacity gain; and (iii) predictive network intelligence leveraging LSTMs to forecast traffic 5 seconds ahead, enabling smarter allocation and reducing congestion by about 60 percent. System-level evaluations combining propagation modeling and validated link-budget analysis demonstrate that this architecture delivers up to an 84 times improvement in useful energy delivery, reduces deployment costs by nearly 74 percent, and eliminates diesel dependence through solar-powered operations, thereby enabling sustainable, green connectivity for both rural and ultra-dense urban environments.

A substantial share of the Earth's land surface is managed by humans, with cities representing the most extreme form of anthropogenic land use. There are zillion ways in which settlements can be arranged across a given area, and their specific spatial configuration has important consequences for both urban systems and the natural environment. Here, we introduce a novel approach to characterizing settlement configuration by systematically quantifying it in terms of a transition resembling percolation -- that is, by identifying the critical distance at which isolated settlements merge into a giant, overarching settlement cluster. We estimate this critical distance across multiple spatial scales and units, including national and subnational levels, non-overlapping tiles, and moving windows, covering the entire globe. The critical distance provides an independent measure of settlement connectivity and thus adds value to spatial analyses of settlement structure and its social, economic, and ecological impacts. Accordingly, our Global Settlement Percolation (GSP) dataset is relevant to a wide range of research communities, including those studying urban morphology, land-use patterns, and landscape ecology.

Network energy efficiency is of critical importance to mobile network operators for economic and ecological reasons. The advent of the O-RAN architecture has brought disaggregation and virtualization, and in order to achieve the highest energy savings gains, we need rigorous measurement, analysis, and modeling of energy consumption at both the component and system levels. However, there remains a lack of publicly-available, quantitative data characterizing the behavior of commercial-grade O-RAN systems. In this white paper, we present a detailed energy-efficiency characterization and modeling of a commercial O-RAN system based on comprehensive power and performance measurements, using a network deployment that faithfully replicates a production O-RAN network deployed by a wireless carrier. The results are drawn from an energy test campaign conducted through a joint collaboration between the Open RAN Center for Integration and Deployment (ORCID) Lab Testing and Evaluation (T&E) Project and the Open Networking Foundation / Rutgers WINLAB Energy Efficiency R&D project. The test environment includes an O-RAN system with an AWS-hosted O-CU, a dedicated-server O-DU, and six high-power, multi-band O-RUs. Our results identify the dominant factors influencing power consumption across the O-RAN stack and quantify energy usage variation under different operational and traffic scenarios. These measurements can be used by operators to parameterize power-consumption models, ultimately supporting data-driven energy optimization and more sustainable operation of commercial O-RAN networks.

Time-triggered messages are of crucial importance in modern communication networks. Offline-generated schedules, which specify start times for periodic messages, enable us to achieve deterministic behavior in critical applications. In automotive and avionics domains, so-called signals (measurements and commands) are periodically generated and communicated (via messages) among sensors, controllers, and actuators. However, the message contains not only the useful signal data, but also necessary metadata, e.g., message ID. Metadata is stored as a header or tail and extends the message size; when the signal is very short (as it often is in applications), sending each in a separate message is inefficient. Thus, several signals are grouped into a single message, depending on their periodicity and length, and sent with just one header. Such an approach increases the utilization of the communication resource (link or bus), since less bandwidth is wasted on headers (Kuaban et al. 2021). However, grouping the signals into messages is complicated. The maximum size of the message (including the metadata) is finite, since longer messages have a lower probability of successful delivery. Also, longer messages are less flexible for scheduling in a periodic setting. This is similar to the work of Huan et al. (2019), where the compromise between energy efficiency and latency for IoT devices was investigated. In this paper, we study the fundamental problem of grouping time-triggered signals into messages and periodic scheduling of messages on a single resource.

Emerging connected-vehicle (CV) data shows great potential in urban traffic monitoring and forecasting. However, prior CV-based studies on arterial traffic measures prediction are limited to simulated high-penetration scenarios or small networks, which are challenging to apply in real-world city-scale arterial networks. To address such gaps, we develop a CV data-based arterial traffic prediction framework with two components: (1) a two-stage traffic state extraction method that estimates vehicle-level traffic measures from CV trajectories and then aggregates them into network-level traffic state measures; (2) an Abnormality-aware spatiotemporal graph convolution network (AASTGCN) that adopts a dual-expert architecture to separately model normal and abnormal traffic, and jointly captures short-term traffic dynamics and long-term periodicity via spatiotemporal GCN with a gated-fusion mechanism. Real-world CV data are used to test our method in a large arterial network with 1,050 links. Experimental results show that: 1) The proposed traffic estimation method is effective for large arterial networks to provide real-time traffic measures (e.g., link-level average travel delay and queue length), which are critical for urban traffic operation and evaluation. 2) Abnormal traffic prediction is typically challenging for existing methods. By modeling abnormal cases separately from normal traffic in two dedicated experts, AASTGCN outperforms existing models for both normal and abnormal traffic conditions. 3) The gate-fusion mechanism adaptively balances real-time and historical information: it leverages more historical-periodic information in normal traffic and shifts a higher weight to real-time traffic dynamics for abnormal traffic deviating abruptly from historical patterns.

Oscillatory dynamics are common features of complex networks, often playing essential roles in regulating function. Across scales from gene regulatory networks to ecosystems, delayed feedback mechanisms are key drivers of system-scale oscillations. The analysis and prediction of such dynamics are highly challenging, however, due to the combination of high-dimensionality, non-linearity and delay. Here, we systematically investigate how structural complexity and delayed feedback jointly induce oscillatory dynamics in complex systems, and introduce an analytic framework comprising theoretical dimension reduction and data-driven prediction. We reveal that oscillations emerge from the interplay of structural complexity and delay, with reduced models uncovering their critical thresholds and showing that greater connectivity lowers the delay required for their onset. Our theory is empirically tested in an experiment on a programmable electronic circuit, where oscillations are observed once structural complexity and feedback delay exceeded the critical thresholds predicted by our theory. Finally, we deploy a reservoir computing pipeline to accurately predict the onset of oscillations directly from timeseries data. Our findings deepen understanding of oscillatory regulation and offer new avenues for predicting dynamics in complex networks.

We show that there is a one-round randomized distributed algorithm that can 2-color cycles such that the expected fraction of monochromatic edges is less than 0.24118. We also show that a one-round algorithm cannot achieve a fraction less than 0.23879. Before this work, the best upper and lower bounds were 0.25 and 0.2. Our proof was largely discovered and developed by large language models, and both the upper and lower bounds have been formalized in Lean 4.

We propose a monolithic, electrically driven source of photon pairs based on a non-linear AlGaAs Bragg reflection waveguide and a laser structure stacked on top. By introducing lateral tapers, the fundamental mode of the lasing waveguide is vertically coupled into a higher order mode of the Bragg reflection waveguide (Bragg mode) such that photon pairs can be generated through a type-II spontaneous parametric down conversion process. According to numerical simulations, a coupling efficiency of 28% is achieved between both modes. Phase matching the Bragg mode with two fundamental modes at 1550 nm results in a photon pair rate of 1.7*10^8 pairs/s for a 2 mm long device assuming 1 mW of power in the Bragg mode. Since the Bragg reflection waveguide does not require doping for this vertically coupled structure, free-carrier absorption losses and parasitic luminescence are avoided.

We present a unified framework for heavy tetraquark and pentaquark systems within the quark-diquark effective mass formalism, extending its baryon-calibrated construction to multiquark states without introducing sector-dependent parameters. Intra-diquark color-spin correlations are encoded in effective diquark masses fixed from baryon spectroscopy, while the inter-cluster chromomagnetic scale, independently determined from vector-pseudoscalar meson splittings, is propagated unchanged to exotic configurations, ensuring residual one-gluon exchange dynamics only between composite color sources. Within this framework, we compute the complete spectra for both $\bar{\mathbf{3}}_c\otimes\mathbf{3}_c$ and $\mathbf{6}_c\otimes\bar{\mathbf{6}}_c$ configurations in tetraquarks, whereas the pentaquark analysis focuses on the dominant $\bar{\mathbf{3}}_c\otimes\bar{\mathbf{3}}_c\otimes\bar{\mathbf{3}}_c$ clustering. Heavy-quark spin symmetry and flavor-symmetry breaking across light, charm, and bottom sectors emerge naturally through the explicit $1/(m_{D_1}m_{D_2})$ scaling of the calibrated couplings. The resulting spectra exhibit a coherent dynamical hierarchy spanning baryons and multiquark states. Established exotic candidates are reproduced within hadronic uncertainties, while the unified calibration enables quantitative predictive control across flavor sectors. The framework thus provides a parameter-economical, systematically constrained baseline with unified dynamical consistency for heavy multiquark spectroscopy.

NRT is an international project to build and operate the world's largest robotic telescope. The telescope will have a segmented primary mirror with an equivalent diameter of 4 m, a set of simultaneously mounted optical and near-infrared instruments, and a response time of less than 30 seconds. The project builds on the experience gained with the successful twenty-year operation of the Liverpool telescope, and with the GTC optics and control system. All of the above together with the excellent conditions for astronomical observation of La Palma, represents a solid base and guarantees that NRT will be one of the leading facilities in the field of time domain astronomy. This contribution will analyze the current status of the project with special emphasis on the development of its optics, and the plans for its construction and operation.

The New Robotic Telescope (NRT) is an international collaboration to build and operate a 4 m diameter fully robotic telescope. The telescope will take advantage of the superb atmospheric conditions at the Observatory of the Roque de los Muchachos (ORM). In conjunction with a large aperture, entirely robotic operation, quick response, and a set of versatile instrumentation in the optical and near-infrared this guarantees a high scientific impact focused mainly in the area of time domain astronomy. This paper presents the scientific motivation and the status of the project, discussing possible technical solutions under evaluation for the optics, mechanics and control system.

Ridges in galaxy density fields measured by photometric surveys are 2D projections of filaments in the cosmic web, and so should lens light from background galaxies. We report on a detection of this effect in Dark Energy Survey Year 3 data at high significance, though not independently of galaxy-galaxy lensing. We describe improvements to the existing subspace-constrained mean shift algorithm to locate these ridges efficiently at scale, and examine the dependence of the signal in simulations on cosmological and algorithmic parameters. We find that it depends primarily on $S_8=σ_8 \left( Ω_m / 0.3 \right)^{1/2}$, and discuss improvements to our methodology that would be needed to allow precision parameter estimation.

Synthetic data is increasingly used to support research without exposing sensitive user content. Social media data is one of the types of datasets that would hugely benefit from representative synthetic equivalents that can be used to bootstrap research and allow reproducibility through data sharing. However, recent studies show that (tabular) synthetic data is not inherently privacy-preserving. Much less is known, however, about the privacy risks of synthetically generated unstructured texts. This work evaluates the privacy of synthetic Instagram posts generated by three state-of-the-art large language models using two prompting strategies. We propose a methodology that quantifies privacy by framing re-identification as an authorship attribution attack. A RoBERTa-large classifier trained on real posts achieved 81\% accuracy in authorship attribution on real data, but only 16.5--29.7\% on synthetic posts, showing reduced, though non-negligible, risk. Fidelity was assessed via text traits, sentiment, topic overlap, and embedding similarity, confirming the expected trade-off: higher fidelity coincides with greater privacy leakage. This work provides a framework for evaluating privacy in synthetic text and demonstrates the privacy--fidelity tension in social media datasets.

We propose to implement an optical analogue of the tautochrone property of the cycloid to allow the focusing of ultrashort pulses inside photonic systems. This allows to enhance nonlinear effects, resulting in orders of magnitude increase of nonlinearity-induced phase shifts, while employing low irradiances. Building upon the optical-mechanical analogy, we show how to produce optical limiters for temporal light pulses, and how to implement temporal bistability and even multistability with large numbers of states. Finally, we move this concept to the quantum realm and predict a tautochrone quantum blockade regime with a stronger antibunching.

Popular social media platforms TikTok, Facebook and Instagram allow third-parties to run targeted advertising campaigns on sensitive attributes in-platform. These ads are interactive by default, meaning users can comment or ``react'' (e.g., ``like'', ``love'') to them. We find that this platform-level design choice creates a privacy loophole such that advertisers can view the profiles of those who interact with their ads, thus identifying individuals that fulfill certain targeting criteria. This behavior is in contradiction to the promises made by the platforms to hide user data from advertisers. We conclude by suggesting design modifications that could provide users with transparency about the consequences of ad interaction to protect against unintentional disclosure.

We present the spatial part of the point source signal extraction strategy for the upcoming CHORD galaxy survey. CHORD, the Canadian Hydrogen Observatory and Radio-transient Detector, is an under-construction drift-scanning compact interferometric radio telescope. CHORD comprises 512 six meter dishes and observes in the 300 to 1500 MHz frequency range. One of its science goals is producing a catalogue of galaxies detected by the neutral hydrogen (HI) 21 cm emission line. CHORD's highly redundant dish layout creates the problem of spatial aliasing, the effect where the same signal could be feasibly produced from sources at multiple locations on the sky. The search will be done with a matched filter in the visibility plane. This paper presents the search strategy and a prediction tool that can quickly estimate the matched filter response at a given sky position, allowing a prediction of alias locations and severity. This tool confirms that although aliases are impossible to distinguish in a single snapshot, they become possible to distinguish when combining data over a period of time. It predicts that aliases will be harder to distinguish for observations closer to the celestial equator, but that scanning with offset adjacent strips can remove this degeneracy. It predicts that the optimal strategy for a single offset to disambiguate aliases is to re-point the array in declination by about two degrees. A future paper will combine these findings with realistic noise estimates and galaxy population statistics to make forecasts of the population of galaxies that CHORD will detect.

We introduce a concept of asymptotic mean of digits (symbols) in the $Q_s$-representation of a real number, that is a generalization of the $s$-adic representation and have a self-similar geometry. We discuss its relationship with the frequencies of digits and formulate problems related to the concept. We study the topological, metric, and fractal properties of the set of real numbers that have no asymptotic mean of $Q_s$-symbols. Also we study topological, metric and fractal properties of the sets of real numbers that have asymptotic mean of $Q_3$-symbols which is equal to value of digit frequency of number.

Density Functional Theory (DFT) is a cornerstone of materials science, yet executing DFT in practice requires coordinating a complex, multi-step workflow. Existing tools and LLM-based solutions automate parts of the steps, but lack support for full workflow automation, diverse task adaptation, and accuracy-cost trade-off optimization in DFT configuration. To this end, we present TritonDFT, a multi-agent framework that enables efficient and accurate DFT execution through an expert-curated, extensible workflow design, Pareto-aware parameter inference, and multi-source knowledge augmentation. We further introduce DFTBench, a benchmark for evaluating the agent's multi-dimensional capabilities, spanning science expertise, trade0off optimization, HPC knowledge, and cost efficiency. TritonDFT provides an open user interface for real-world usage. Our website is at https://www.tritondft.com. Our source code and benchmark suite are available at https://github.com/Leo9660/TritonDFT.git.

Reasoning is the ability to integrate internal states and external inputs in a meaningful and semantically consistent flow. Contemporary machine learning (ML) systems increasingly rely on such sequential reasoning, from language understanding to multi-modal generation, often operating over dictionaries of prototypical patterns reminiscent of associative memory models. Understanding retrieval and sequentiality in associative memory models provides a powerful bridge to gain insight into ML reasoning. While the static retrieval properties of associative memory models are well understood, the theoretical foundations of sequential retrieval and multi-memory integration remain limited, with existing studies largely relying on numerical evidence. This work develops a dynamical theory of sequential reasoning in Hopfield networks. We consider the recently proposed input-driven plasticity (IDP) Hopfield network and analyze a two-timescale architecture coupling fast associative retrieval with slow reasoning dynamics. We derive explicit conditions for self-sustained memory transitions, including gain thresholds, escape times, and collapse regimes. Together, these results provide a principled mathematical account of sequentiality in associative memory models, bridging classical Hopfield dynamics and modern reasoning architectures.