arXiv subjects

Tong Liu

Publications and source records attributed to Tong Liu.

At least 19 recordsLinked to original sources

arXiv2026

Migration Traps as Variability Attractors. II. Nonmonotonic X-Ray Timing Signatures of Embedded Stellar-Mass Black Holes in AGN Disks

Stellar-mass black holes (sBHs) embedded in active galactic nucleus (AGN) disks may affect X-ray variability through magnetic reconnection. As sBHs migrate and accumulate near migration traps, they disturb the disk magnetic field and trigger reconnection events that heat the disk and eject plasma into the overlying corona. The injected plasma dissipates kinetic energy and can trigger the release of magnetic energy already stored in the corona. We calculate the resulting coronal X-ray emission and the disk ultraviolet (UV) variability, including the disk thermal response and reprocessing of incident X-rays. We find that embedded sBHs can initially enhance short-timescale X-ray variability, but the variability becomes suppressed as more sBHs accumulate in the disk. At higher mass accretion rates through AGN disks, this suppression appears earlier. As sBH-associated reconnection events become more frequent, their active regions increasingly overlap, reducing the magnetic energy available to each event from the corona. We also find that variations in the disk ultraviolet (UV) emission follow the X-ray variations with a shorter delay than in the model without sBHs. These results suggest that a growing population of embedded sBHs does not necessarily produce stronger short-timescale X-ray variability.

astro-ph.HE↗

arXiv2026

LLMs Can Tackle High Energy Theoretical Physics, yet Gaps Remain: Part I, Quantum Field Theory

Beyond math and code, Large language models (LLMs) are beginning to tackle problems that have traditionally required substantial training for theoretical physicists, but it remains unclear how far this capability extends and where it might break. As a first step toward pushing LLMs in high-energy theoretical physics, we introduce QFT-Bench, a benchmark of 75 graduate- to research-level problems requiring extensive reasoning and derivation across core Quantum Field Theory (QFT) subareas. Each question has a well-defined target and closed-form reference answer. Evaluation combines symbolic verification with LLM-based judge. Across 39 proprietary and open-source models, frontier models already demonstrate substantial capabilities on nontrivial QFT calculations, with the strongest models exceeding 95% accuracy. At the same time, clear gaps remain. Even strong models still make errors in calculations, conventions, and instruction following, while weaker models often fail at the level of physical concepts or derivations. We further monitor models' CoTs and show that current models exhibit nontrivial reasoning unfaithfulness, i.e., they could reach correct answers but the reasoning itself does not justify and logically support them. These results provide a quantitative assessment of current LLM capabilities in QFT and a first step toward broader evaluation in high-energy theoretical physics.

hep-ph↗

arXiv2026

A Full-Spectrum Theory of Moiré Superlattices

Conventional theoretical descriptions of moiré superlattices, epitomized by the Bistritzer--MacDonald continuum model, are fundamentally constrained by low-energy Dirac cone approximations and first-shell interlayer coupling truncations. While successful for small-angle twisted bilayer graphene near charge neutrality, these frameworks fail to provide universal analytical solutions for strained lattices, high-energy band behaviors, or arbitrary commensurate geometries. Here, we establish a unified, approximation-free analytical formalism for generalized moiré systems based on three pillars: (i) a rigorous two-dimensional-to-one-dimensional supercell cyclic mapping, (ii) exact block-circulant diagonalization of intralayer Hamiltonians, and (iii) quasi-degenerate Löwdin perturbation theory centered on arbitrary energy degeneracy points across the full spectrum. Unlike traditional schemes anchored to the Dirac point, our approach enables systematic perturbative expansions around any band crossing, unifying the characterization of low-energy flat bands with high-energy degeneracies. We demonstrate that this framework perfectly recovers the Bistritzer--MacDonald model in the low-energy limit while extending its reach to heterostrained systems and high-energy regimes. By incorporating high-order interlayer coupling shells, we reveal an intrinsic scale self-similarity in moiré lattices and predict the emergence of higher-order magic angles and tunable flat bands driven by secondary degeneracies. This work eliminates the long-standing lack of universal analytical tools for diverse moiré platforms, providing a full-spectrum theoretical foundation for band engineering and the exploration of exotic correlated and topological states in van der Waals heterostructures.

cond-mat.mes-hall↗

arXiv2026

Beam Spill Structure Measurements at CEE Using the T0 Detector

The Cooling Storage Ring External-target Experiment (CEE) at the Heavy Ion Research Facility in Lanzhou uses a slow-extracted heavy-ion beam whose temporal profile directly affects detector occupancy. The existing CEE T0 detector, originally designed to provide the event start time, is extended in this work to measure the spill structure immediately upstream of the fixed target. The detector consists of a 0.8-mm-thick EJ-200 plastic scintillator read out by 16 silicon photomultipliers at each end. The SiPM signals are conditioned and discriminated by the NINO front-end electronics, and the resulting timing signals are recorded using an FPGA-based time-to-digital converter. In laboratory tests, with the laser trigger set to its maximum available repetition frequency of 100~MHz, the recorded T0 hit-rate distribution was concentrated around 100~MHz. During the June 2024 beam test with 350-MeV/u Kr ions, the accelerator team varied several operating settings, including RF-KO, sextupole, and feedback settings, while T0 independently recorded the corresponding spill profiles. An empirical representation was used to extract the rise, plateau, and fall durations from the stored profiles. The plateau duration was used as the primary comparison metric while accelerator stability and RF-system load were also considered. The measurements clearly distinguished the spill-profile responses associated with different accelerator settings. The spill obtained with the selected settings exhibited a more nearly rectangular envelope than the initial profile, although residual fluctuations and a terminal spike remained. These results demonstrate that the existing CEE T0 detector can provide independent beam-structure information during accelerator tuning.

physics.ins-det↗

arXiv2026

Reference-Free Singing Pitch Correction via Music-Constrained Sequence Editing

Existing singing pitch correction approaches rely on target melodies or accompaniment tracks, which may be unavailable in practice. We formulate reference-free singing pitch correction as a music-constrained sequence editing task that determines whether and how each note should be corrected from the input performance alone. A pretrained symbolic music encoder with lightweight singing-domain adapters produces contextual representations of the singing MIDI. Based on these representations, two lightweight correction heads jointly model correction necessity and signed pitch modification through a factorized pitch-editing distribution. An input-dependent tonal prior derived from the estimated key distribution then reranks the candidate offsets, favoring tonally compatible corrections without target melodies or ground-truth key annotations. Experiments on real paired amateur and professional singing recordings show that the method improves note-level pitch accuracy from 73.25 to 83.43, outperforming a context-based baseline by 3.99 percentage points while balancing error correction and preservation of correctly performed notes.

cs.SD↗

arXiv2026

Breaking Fault Lines: Unifying TEE-Assisted BFT Consensus in Partially Trusted Worlds

This paper revisits TEE-assisted BFT under a universal partial-TEE model, where an arbitrary subset of replicas execute inside TEEs while the remaining replicas operate without hardware trust guarantees. We show that heterogeneous trust changes the structure of quorum formation and fault tolerance. In particular, we derive a tight resilience bound f < max {n/3, m/2}, where n is the total number of replicas and m is the number of TEE-enabled replicas. The result reveals a sharp threshold phenomenon: TEEs improve fault tolerance only once they exceed two-thirds of the deployment. Guided by this characterization, we introduce two protocol principles: (1) a dual-quorum construction that safely combines TEE-only and mixed quorums, and (2) a TEE-leader fast path that leverages hardware-enforced non-equivocation to reduce both consensus and view-change latency. We realize these ideas in Raftel, which is, to our knowledge, the first HotStuff-style BFT protocol designed explicitly for arbitrary partial-TEE deployments, and in chained-Raftel, a pipelined variant that further accelerates mixed-trust execution. We implement both protocols atop Intel SGX and evaluate them in LAN and WAN environments. Our results show that Raftel achieves up to 625 TPS with sub-670 ms latency in WAN settings, outperforming HotStuff by up to 308 TPS in throughput while approaching the performance of fully TEE-assisted protocols.

cs.DC↗

arXiv2026

EEGDM: Learning EEG Representation with Latent Diffusion Model

Recent advances in self-supervised learning for EEG representation have largely relied on masked reconstruction, where models are trained to recover randomly masked signal segments. While effective at modeling local dependencies, the training objective of masked reconstruction does not compel the model to capture global generative constraints essential for characterizing neural activity. To address this limitation, we propose EEGDM, a novel self-supervised framework that leverages latent diffusion models to generate EEG signals as an objective. Unlike masked reconstruction, diffusion-based generation progressively denoises signals from noise to realism, compelling the model to capture holistic temporal patterns and cross-channel relationships. Specifically, EEGDM incorporates an EEG encoder that distills raw signals and their channel augmentations into a compact representation, which serves as conditional information to guide the diffusion denoising process, thereby enabling the encoder and diffusion model to be jointly optimized through the generative objective. This design endows EEGDM with a compact latent space, which not only offers ample control over the generative process but also can be leveraged for downstream tasks. Experimental results show that EEGDM (1) reconstructs high-quality EEG signals, (2) learns robust representations, and (3) achieves competitive performance across diverse downstream tasks, thus exploring a new direction for self-supervised EEG representation learning.

cs.LG↗

arXiv2026

RICE-Alpha: Reliability-Informed Correction with Event Graphs for LLM-Agent Stock Forecasting

Equity-relevant news evolves through temporally dependent corporate events, making historical information useful only when event continuity, information availability, and transition reliability are modeled. Existing LLM-based financial agents incorporate historical evidence, yet they provide limited support for preserving issuer-specific chronology under point-in-time constraints and for identifying when historical transitions contribute information beyond the current forecast. We present RICE-Alpha (Reliability-Informed Correction with Event Graphs), a point-in-time stock-scoring framework that separates a history-aware multi-view Base Alpha from a reliability-calibrated residual correction derived from historical event continuation. A Multi-Tier Memory Layer grounds news interpretation in temporally eligible issuer-specific history, while a Typed Event Agent constructs event states whose successor relations are formed within issuers and pooled across firms only after valid local pairing. Matured transitions are calibrated by their empirical reliability, and the resulting graph signal is residualized against the Base Alpha and technical view to obtain the RICE Delta. On daily Nasdaq-100 and Hang Seng Index panels from 2024 to 2026, RICE-Alpha achieves the strongest results among the evaluated LLM-based agents and momentum across four predictive and four portfolio-level metrics. Its ICIR more than doubles that of the strongest baseline, while net Sharpe ratios reach 1.656 and 1.725 in the U.S. and Hong Kong, respectively. U.S. ablations further show significant reductions in IC and RankIC after Holm adjustment when major components are removed. These results indicate that historical event continuation adds incremental information when it is temporally grounded, reliability-calibrated, and introduced as a residual correction to a multi-view forecast.

cs.LG↗

arXiv2026

Rethinking Multi-Image Re-Representation in Multi-Image Understanding

Multi-image understanding requires MLLMs not only to recognise the content of individual images, but also to organise visual evidence distributed across them. We study this problem through multi-image re-representation, viewing prompted Chain-of-Thought reasoning and agentic visual tool use as different ways of re-organising visual evidence during reasoning. We introduce Mosaic, a general-purpose multi-image visual harness that enables an MLLM to actively construct visual intermediates with ten composable image operations. We compare five re-representation settings on existing multi-image benchmarks and on MosaicBench, a new grounding-focused benchmark for fine-grained multi-image understanding. Our experiments show that the relative benefits of textual and visual re-representation are strongly task-dependent. Visual re-representation is particularly effective for tasks requiring precise visual evidence, including hypothesis testing, precision comparison, and orientation-sensitive reasoning, while tasks dominated by higher-level semantic content show smaller or less consistent gains. Building on this finding, we train MosaicAgent-8B to use Mosaic with reinforcement learning using only accuracy and format rewards. Without demonstration trajectories or rewards for specific tool-use, the agent learns to compose visual operations over multiple steps and exhibits diverse problem-solving patterns unpromptedly. Code and data will be released at https://github.com/gengyuanmax/Mosaic.

cs.CV↗

arXiv2026

Recurrent double-peaked $γ$-ray sub-flares in PKS 1424-418

We report a hint of recurrent double-peaked $γ$-ray sub-flares in PKS~1424--418 during two similar major active epochs, MJD 56117--56498 and MJD 59669--59978. The candidate double-peaked sub-flares show a characteristic intra-subflare peak separation of $\sim 11.8$ d and a neighbouring-subflare spacing of $\sim 67$ d. Red-noise Monte Carlo tests indicate that the candidate recurrent double-peaked morphology is not easily reproduced by stochastic variability alone, with false-alarm probabilities of $p_{\rm a}=0.0216$ and $p_{\rm b}=0.0006$ for the two active epochs, respectively. In the simulations of the entire \textit{Fermi} Large Area Telescope (LAT) light curve, no red-noise realisation reproduces both epoch-like structures. The spectral energy distribution (SED) modelling shows similar radiative properties for the two peaks of one double-peaked candidate, while changes in the Doppler factor appear to play an important role in the flaring activity. Together with the previously reported near-zero lag between the millimetre and $γ$-ray emission, these results suggest that the recurrent activity may be associated with the compact mm/sub-mm core region. We discuss a possible structured-jet scenario in which a single disturbance propagates through a chain of quasi-stationary recollimation shocks. In this picture, repeated interactions between the disturbance and the shock structure may qualitatively account for both the double-peaked sub-flare profiles and the observed $\sim 11.8$ d and $\sim 67$ d time-scales. Further active epochs are required to determine whether this similarity represents a recurrent long-term behaviour.

astro-ph.HE↗

arXiv2026

Migration Traps as Variability Attractors. I. Optical/UV Signatures of Embedded Stellar-Mass Black Holes in Active Galactic Nucleus Disks

We investigate whether embedded stellar-mass black holes (sBHs) in active galactic nucleus (AGN) disks can leave observable optical/UV variability signatures through migration-trap-driven magnetic heating. This mechanism operates when sBHs migrating toward torque-balance radii pile up near migration traps, triggering localized, stochastic magnetic reconnection that heats the disk atmosphere. It is potentially important because it provides a physical source of non-coronal disk heating and directly links optical/UV continuum variability to otherwise hidden compact-object populations. By coupling a one-dimensional sBH population synthesis model with a corona-heated accretion-disk reprocessing variability framework, we show that migration traps concentrate sBHs at preferred radii and generate localized, stochastic reconnection heating. The resulting heating is self-regulated: sBH pile-ups enhance the reconnection rate, while gap opening reduces the local gas density and partially suppresses the reconnection power. This heating produces excess short-timescale optical/UV variability, flattened short-term structure functions, and deviations from the standard $τ\proptoλ^{4/3}$ lag-wavelength relation, which describes the time delay between variability at different wavelengths for a standard thin accretion disk. These signatures are strongest at low-to-moderate Eddington ratios, and related observations could provide indirect evidence for embedded compact-object populations in AGN disks.

astro-ph.GA↗

arXiv2026

Quasi-periodic Eruptions from Recurrent Satellite Black Hole Transits through Magnetized Galactic Nucleus Accretion Disks

Quasi-periodic eruptions (QPEs) are recurrent soft X-ray flares from galactic nuclei, but their origin remains uncertain. The delayed ultraviolet (UV) counterpart detected in ZTF19acnskyy provides a new constraint on viable models. We present a two-channel model in which a satellite black hole (sBH) repeatedly crosses a nuclear accretion disk threaded by a large-scale magnetic field. Gravitational focusing and dynamical drag generate hot, optically thick ejecta whose expansion and photon diffusion power the soft X-ray QPE. For fiducial Bondi-scale parameters, the model yields a characteristic X-ray duration of $\sim10^3\ \mathrm{s}$ and luminosity of $\sim10^{42}\ \mathrm{erg\,s^{-1}}$. For ZTF19acnskyy, the model reproduces the observed day-scale X-ray duration and energetics. Simultaneously, the sBH motion compresses and bends the background magnetic field, triggering in-disk reconnection. The reconnection channel provides the energy budget and photon-diffusion delay required for the variable UV component. Unfavorable magnetic fields or diffusion times longer than the QPE recurrence period can weaken or smear out the UV signal, potentially explaining the lack of clear UV counterparts in other QPE sources.

astro-ph.HE↗

arXiv2026

Atria Dawn: The Dawn of Agentic Superintelligence

As AI agents become participants in the development of their successors, they reshape both the production of intelligence and the role of human researchers. We introduce Atria Dawn Preview, a foundation agentic language model designed for scientific research and engineering workflows, with the goal of expanding the frontier of agent productivity in the real world. This model is trained via a Verifiable Experience Pipeline that connects tool-mediated interactions to executable environments and externally verified outcomes. Across 16 benchmarks spanning real-world research, engineering, and digital work, Atria Dawn Preview is competitive with frontier agents and achieves the highest reported score on five of them. Beyond standalone performance, we examine the real research-and-development process behind this model as a case study of human--AI collaboration, analyzing 769 task records from 56 participants together with agent logs. When asked to evaluate completed tasks under comparable conditions, participants rated about one-third of completed AI-assisted tasks as infeasible without AI. More strikingly, agents frequently propose methods and implement revisions, while humans retain most final decisions and guide exploration through judgment and feedback. These observations indicate a shift from task-level execution to project-level partnership, with human effort concentrating on what is worth pursuing and how evidence should guide research. Progress toward more autonomous AI research must therefore advance both the capacity for discovery and the capacity for meaningful human oversight, preserving accountable human authority over the risks and direction of continued development.

cs.AI↗

arXiv2026

Little Red Dots as Supermassive Counterparts of SS 433: A Hyper-Eddington Accretion Framework

High-redshift little red dots (LRDs) are compact sources characterized by V-shaped spectral energy distributions (SEDs), broad emission lines, and often prominent Balmer breaks. Their high number density and apparently large black hole masses suggest that they are essential to the early evolution of galaxies and supermassive black holes (SMBHs); however, the nature of their central engines remains uncertain. Here, we propose that LRDs are the supermassive, high-redshift counterparts of the hyper-Eddington accreting Galactic microquasar SS~433, viewed at high inclinations. By scaling the hyper-Eddington accretion physics from stellar-mass black holes to supermassive scales, we show that the observed LRD features, including X-ray weakness, soft optical SEDs, apparent sub-Eddington accretion ratio, and Balmer breaks, emerge naturally from the self-shielding geometry of a puffed-up accretion disk. In this framework, the broad-line regions are ionized by anisotropic radiation escaping from the inner disk, analogous to the unseen UV/X-ray emission revealed by the W50 nebula in SS 433. Their low-inclination or lower-accretion-rate counterparts would appear as little blue dots (LBDs) or normal active galactic nuclei. Our model predicts that the Balmer break strength positively correlates with the broad-line width, that the emission lines are more variable than the optical continuum, that LRDs are intrinsically more luminous than observed, and that LBDs are more variable than LRDs. This unified-scale model redefines LRDs as the essential laboratories for observing the rapid accretion-driven growth that shaped the early assembly of galaxies and their central SMBHs.

astro-ph.HE↗

arXiv2026

Deep Research Agents Brings Deeper Harm

We reveal that Deep Research (DR) agents systematically expose safety risks: simply submitting harmful queries that a standalone LLM would reject outright can elicit detailed and dangerous reports from DR agents. Empirical analysis reveals that the advantages that make DR agents powerful unintentionally make them vulnerable: both the research role assignment (e.g., Planner) and the multi-step execution mechanism weaken alignment of the base LLM, leading to severe safety breaches. Through linear probe analysis of internal hidden states, we demonstrate that role assignment suppresses refusal awareness by shifting representations away from safety boundaries. Besides, multi-step execution distributes harmfulness across individual steps, preventing alignment mechanisms from being activated throughout the research process. Exploiting these vulnerabilities, we design Intent Hijack (i.e., rephrasing harmful queries as academic research) and Plan Injection (i.e., manipulating execution plans) to further examine the safety risks of DR agents. Extensive experiments show that our methods achieve near-perfect compliance and elicit detailed, actionable reports that significantly exceed standalone LLM outputs in technical depth and applicability. These results demonstrate alarming misalignment in DR agents and underscore the urgent need for tailored alignment techniques. Our code is available in https://schen.app/deeper-harm.

cs.CR↗

arXiv2026

Tunable topological enhancement of covariant quantum Fisher information via non-Bloch skin effect in non-Hermitian SSH lattices

The covariant quantum Fisher information (CQFI) has recently been established as the ultimate precision benchmark for pseudo-Hermitian sensors [Phys. Rev. Lett. 136, 080802 (2026)], yet existing analyses are limited to single-mode systems. Here we extend the CQFI formalism to multi-mode non-Hermitian Su-Schrieffer-Heeger lattices and reveal tunable topological enhancement enabled by the non-Hermitian skin effect (NHSE). Under open boundary conditions, the NHSE deforms the conventional Brillouin zone into a generalized Brillouin zone of radius r = exp(\k{appa}), where \k{appa} denotes the non-Bloch decay rate. While the total CQFI scales linearly with system size N, its prefactor depends critically on \k{appa}, yielding an enhancement factor E(N) = F_OBC / F_PBC that exceeds 30 for N ~ 30, substantially outperforming periodic-boundary sensors. The enhancement is robust against moderate local disorder and supports multi-parameter estimation, with the joint Cramér-Rao bound reduced by up to 15 orders of magnitude. These findings establish a spatial-domain mechanism for quantum metrology that complements time-domain strategies and is experimentally accessible using topoelectrical circuits, photonic lattices, and superconducting circuits with current technology.

quant-ph↗

arXiv2026

When Compression Helps and When It Hurts: Condition-Aware Analysis of Chain-of-Thought Distillation

Chain-of-Thought (CoT) distillation transfers multi-step reasoning from large reasoning models to smaller students, but verbose teacher traces inflate both training and inference cost. Existing CoT compression methods fall into two families, selective pruning and generative rewriting, yet prior studies have left key factors entangled: granularity is confounded with importance criteria in pruning, restructuring level is rarely isolated in rewriting, and compression budgets are not systematically evaluated across domains or regimes. We recast CoT compression along three dimensions: importance criterion, restructuring level, and compression budget. Sweeping these across two model families, Math and General domains, and Long-/Short-CoT regimes, we find that (i) importance criterion utility is strictly governed by granularity: step-level criteria converge on a shared reasoning backbone, while token-level pruning requires symbol-aware signals to preserve the logical core; (ii) restructuring level inverts across domains: Math degrades monotonically with structural disruption, while aggressive rewriting acts as a denoiser on General tasks; (iii) training-time compression does not necessarily translate to inference-time savings: Long-CoT students retain verbose habits despite concise supervision, making the training ratio an optimistic lower bound on deployment cost. These findings yield condition-aware guidelines for matching compression to deployment context.

cs.CL↗

arXiv2026

Reliable LLM-Generated Programs for High-Energy Physics Experiments through Graph-Grounded Software Knowledge

Extracting physics information from modern particle-physics experiments requires multistage analyses implemented on top of large and highly interconnected software ecosystems. General-purpose large language models (LLMs) often produce unreliable programs for such tasks because a user request alone rarely specifies the required APIs, dependencies, and usage conventions. We organize these software relations before generation and retrieve task-relevant knowledge at inference time. Using the open-source ROOT framework as a representative and reproducible testbed, we evaluate a complete grounding system that combines hybrid retrieval over a heterogeneous software knowledge graph, skill-selected workflow examples, and execution-guided repair. On a benchmark of 275 ROOT tasks, grounding improves first-attempt execution from 58.5% to 76.0% under Claude Code orchestration and from 51.3% to 64.0% under standalone orchestration. Final success increases from 90.5% to 96.0% and from 78.9% to 90.9%, respectively, while the average generation cost per successful task increases by only 1.3% and 3.2%. The gains persist under a strong coding agent, indicating that explicit software knowledge remains valuable even when agentic scaffolding is already in place. Because the method captures software relations common to large codebases rather than facts specific to ROOT or a particular model, it should transfer to other experiment frameworks and proprietary software, especially where documentation is sparse or internal dependencies are complex.

cs.SE↗