arXiv subjects

Li Gan

Publications and source records attributed to Li Gan.

15 recordsLinked to original sources

arXiv2026

Prizes, Shared Hazard, and the Speed of Frontier AI

Frontier AI labs compete in a contest for commercial leadership while sharing the risk of catastrophe. Each lab gains from arriving earlier but bears only part of the hazard its speed adds, which this paper calls the hazard externality. In an illustrative calibrated takeoff model, three racing labs get through the takeoff period in 18 months, while a common rule on research shares would stretch it to 69 months. Neither force alone explains racing that fast. Removing the contest first closes over ninety percent of the gap in catastrophe probability between racing and a common rule when the catastrophe probability at full speed is 75 percent, and 69 percent when it is 25 percent; averaged over both orders of removal, the contest explains 65 and 44 percent. More labs make each lab race harder, when industry hazard at a common research share does not depend on the number of labs. Under open source in its ideal form, where every lab publishes its capability findings at once, labs slow to the coordinated pace, because no lab can get ahead and each bears all the hazard its research adds. One lab's restraint achieves much less than a common rule. The White House Accord on Super Intelligence of September 2026 audits safety but sets no limit on speed. In the model it does not slow the race, because each lab still gains as much from speeding up.

econ.GN↗

arXiv2026

Agreements Have a Clock: Compute Growth and the Verification of AI Restraint

Rival powers have agreed to limit nuclear arms, phase out ozone-depleting chemicals, and cut greenhouse gases, though each agreement was hard to reach. This paper asks what an agreement to pause frontier AI needs. A follower advances through own research and diffusion from the leader. Because diffusion fades as the gap closes, a paused leader stays ahead longer than its raw lead suggests. A solved agreement game adds voluntary entry, costly monitoring, secret advances, and credible responses. Cooperation is supportable when the monitoring needed to deter a breakout costs no more than either country gains from cooperating. As compute grows during a pause, a secret advance reaches the frozen frontier sooner, shrinking the response window: agreements have a clock. In an example with a fixed leader window, cooperation is supportable through month 18.9. Verified matched safety research extends this to month 25.5, and halving the monitoring cost to month 31.8.

econ.GN↗

arXiv2026

Optimal Order Placement under a State-Dependent Hawkes Process Limit Order Book model

We study optimal quote placement for a market maker facing order flow modeled by a state-dependent Hawkes process. A block-based linear demand specification links market-order counts to execution volumes, providing the demand moments required by a continuous time discrete state dynamic programming formulation. Under a quadratic terminal inventory penalty, we characterize inventory-dependent quotes and derive backward recursions for the coefficients of a quadratic value function. Numerical experiments using LOBSTER data examine quote distributions, inventory exposure, and terminal wealth. The implemented strategy yields higher aggregate terminal wealth than fixed-spread benchmarks at one and three ticks, but lower wealth than the two-tick benchmark. These results illustrate the framework's numerical implementation and distinguish terminal-wealth comparisons from evaluation under the inventory-penalized objective.

q-fin.TR↗

arXiv2026

Execution and Evaluation: Task Content, Labor Demand, and the Evaluator Pipeline

Measures of occupational AI exposure ask whether machines can perform a task, not which function the human supplies. I distinguish execution, producing a first-pass output, from evaluation, judging whether it is correct, and make the distinction measurable by scoring all 19,265 O*NET task statements under fixed rubrics. The execution score is reproducible across model coders and O*NET vintages. In a 200-task random audit it correlates 0.81 with the mean of four blinded human coders. It predicts each held-out coder's peers better than that coder does, though the coders disagree substantially among themselves. It adds only modest power beyond O*NET's evaluation activities. Employment growth is lower in execution-heavy white-collar occupations in every window since 2012; the gradient shows no post-2022 break and lies largely between occupational families. The Eloundou exposure gradient is steeper in 2022--2025 than in the pandemic window, but the design does not establish an AI effect. A near-term seniority test is null. If evaluative judgment is itself produced by execution practice, the supply of evaluators falls only after a cohort lag; that prediction remains untested.

econ.GN↗

arXiv2026

Optimal Size in Market Making under State-Dependent Hawkes Process

We study market making with a parametric model of executed volume that accounts for quote placement and state-dependent order flow. Executed volume is expressed as the product of an average order size, an exponential demand function, and a Prony-series multiplier depending on quote depth, market-order count constructed from a state-dependent Hawkes model. An approximation result for finite Prony sums on a compact state domain provides a mathematical basis for the representation. An empirical study of limit-order-book data examines the calibrated demand and intensity models, sensitivity to inventory restrictions, and the resulting placement policies. In retrospective evaluations conditional on realised order-flow paths, the selected calibration achieves higher terminal utility than a baseline while respecting the prescribed inventory bound. The comparisons also distinguish the contribution of trading revenues from that of inventory control. The resulting closed-form volume specification takes placement levels as inputs and can therefore be used with alternative quoting rules.

q-fin.TR↗

arXiv2026

Two Types of Tertiarization: Household Demand, Production Networks, and the Rise of Services

I measure whether the rise of services is ultimately supported by household consumption or by non-household demand through production networks, investment, exports, and government. Using harmonized input-output tables for 76 economies over 1995-2022, I attribute service value added to its ultimate domestic final-demand category and measure the household orientation of each service economy in both levels and growth. The conventional service share does not reveal the type: it is negatively correlated with household orientation, and only about four in ten tertiarization episodes are majority-household. Among the 65 economies whose service share rose over 1995-2018, 77 percent became less household-oriented; the share is 75 percent under a two-percentage-point material-change threshold. Yet household-supported production becomes more service-intensive with income because households' aggregate position falls faster. A one-type benchmark has no purchaser margin; a two-type extension gives the condition for this reversal. An exact Leontief accounting and Shapley decomposition allocate a median 73 percent of service-share growth to production-system channels, which are the largest contribution in 80 percent of economies with materially rising shares. The decline in household orientation is instead allocated mainly to aggregate final-demand structure. The service share records how much service activity exists, not which economic transformation produced it.

econ.GN↗

arXiv2026

Risk Ceilings and Development Deadlines: Pacing AI under Uncertain Safety Productivity

Can a regulator promise both a risk ceiling and a development deadline when safety productivity is unknown? A ceiling below the final model's unprotected hazard requires a minimum stock of safety knowledge, so both promises hold only if weak research can be ruled out. Learning first and then replaying development, with spare compute in safety, comes close to that minimum. In a calibration with only state risk, constant safety yield, and full knowledge transfer, it needs only 3.3 percent more productivity than the necessary bound. Customer services pay for the guarantee. Rules that fix compute allocation fix dates, not risk.

econ.GN↗

arXiv2026

Safety Research and the Allocation of Frontier Compute

Who should give up compute for AI safety research: customers or capability research? Safety paid for by customers delays no capability. Safety paid for by research slows the frontier, which gives knowledge more time to accumulate but postpones what a finished frontier provides. In a frontier-takeoff model, the cheaper source at a given allocation is set by a threshold in demand capacity. Demand capacity is the share of compute customers can use once the frontier is finished. Research should pay too only when demand capacity lies below the threshold. Where the threshold sits depends on how fast the value of customer services diminishes. In the calibration it rises with the service budget, because each further unit taken from customers costs them more. Safety paid for by customers should come early: its ratio to customer compute falls over the takeoff unless safety productivity rises fast. In a calibration with a tight service budget, early customer-funded safety lowers the probability of catastrophe from 75 to 51 percent. That is within one point of the best rule found that may also slow research.

econ.GN↗

arXiv2026

Maintaining Human Verification Capacity under Automation

Human verification depends on expertise that must be maintained before it is needed. This paper links reliance on automated checks, investment in human checking ability, and performance during an interruption. Better checking lowers the error reduction gained from an extra unit of human skill while the checker works. It can therefore reduce the incentive to preserve independent expertise, even when it lowers the best achievable expected cost of maintenance and errors. In an illustration, a more informative checker raises detection while it works from \hoPowWorkA{} to \hoPowWorkC{} percent, but the organization then keeps no routine practice and detects \hoPowOnsetC{} percent of errors when the checker first fails, against \hoPowOnsetA{} percent with a less informative checker. A detection requirement therefore concerns both current capability and its survival until new training becomes effective. Evidence from colonoscopy and aviation documents weaker unaided performance under routine automation, without isolating the mechanism. The framework connects a detection target to an explicit reserve of expertise and a training pipeline. Standard detection tests estimate each quantity and reveal automation bias and silent checker failures. The framework distinguishes the requirement from minimizing expected loss and proposes a longitudinal test.

econ.GN↗

arXiv2025

RoundaboutHD: High-Resolution Real-World Urban Environment Benchmark for Multi-Camera Vehicle Tracking

The multi-camera vehicle tracking (MCVT) framework holds significant potential for smart city applications, including anomaly detection, traffic density estimation, and suspect vehicle tracking. However, current publicly available datasets exhibit limitations, such as overly simplistic scenarios, low-resolution footage, and insufficiently diverse conditions, creating a considerable gap between academic research and real-world scenario. To fill this gap, we introduce RoundaboutHD, a comprehensive, high-resolution multi-camera vehicle tracking benchmark dataset specifically designed to represent real-world roundabout scenarios. RoundaboutHD provides a total of 40 minutes of labelled video footage captured by four non-overlapping, high-resolution (4K resolution, 15 fps) cameras. In total, 512 unique vehicle identities are annotated across different camera views, offering rich cross-camera association data. RoundaboutHD offers temporal consistency video footage and enhanced challenges, including increased occlusions and nonlinear movement inside the roundabout. In addition to the full MCVT dataset, several subsets are also available for object detection, single camera tracking, and image-based vehicle re-identification (ReID) tasks. Vehicle model information and camera modelling/ geometry information are also included to support further analysis. We provide baseline results for vehicle detection, single-camera tracking, image-based vehicle re-identification, and multi-camera tracking. The dataset and the evaluation code are publicly available at: https://github.com/siri-rouser/RoundaboutHD.git

cs.CV↗

arXiv2025

Lattice random walks and quantum A-period conjecture

We derive explicit closed-form expressions for the generating function $C_N(A)$, which enumerates classical closed random walks on square and triangular lattices with $N$ steps and a signed area $A$, characterized by the number of moves in each hopping direction. This enumeration problem is mapped to the trace of powers of anisotropic Hofstadter-like Hamiltonian and is connected to the cluster coefficients of exclusion particles: exclusion strength parameter $g = 2$ for square lattice walks, and a mixture of $g = 1$ and $g = 2$ for triangular lattice walks. By leveraging the intrinsic link between the Hofstadter model and high energy physics, we propose a conjecture connecting the above signed area enumeration $C_N(A)$ in statistical mechanics to the quantum A-period of associated toric Calabi-Yau threefold in topological string theory: square lattice walks correspond to local $\mathbb{F}_0$ geometry, while triangular lattice walks are associated with local $\mathcal{B}_3$.

math-ph↗

arXiv2025

Quantum Charging Advantage from Multipartite Entanglement

Collective quantum batteries (QBs) demonstrate remarkable acceleration in charging dynamics compared to their individual counterparts, underscoring the pivotal contribution of quantum correlations to advanced energy storage paradigms. A fundamental challenge lies in identifying QBs that exhibit genuine quantum advantages derived from multipartite entanglement. In this Letter, based on numerical and analytical evidence, we conjecture a universal bound on the charging rate for fully charging schemes, which is determined by the maximum entanglement depth arising during the charging dynamics. Here, the charging rate quantifies the intrinsic evolution speed of the charging process, appropriately normalized against the quantum speed limit (QSL). We analytically validate this conjecture in three distinct scenarios: (i) fully charging schemes saturating the QSL, (ii) fully parallel charging schemes, and (iii) the SU(2) fully charging schemes. Moreover, we establish a novel lower bound for entanglement depth detection, facilitating numerical verification of our proposed conjecture. By defining the genuine quantum charging advantage as the ratio between entanglement-enhanced charging rates and the maximum achievable non-entangling charging rate, we demonstrate that the charging rate constitutes a robust indicator of genuine quantum advantages.

quant-ph↗

arXiv2023

On the algebraic area of cubic lattice walks

We obtain an explicit formula to enumerate closed random walks on a cubic lattice with a specified length and 3D algebraic area. The 3D algebraic area is defined as the sum of algebraic areas obtained from the walk's projection onto the three Cartesian planes. This enumeration formula can be mapped onto the cluster coefficients of three types of particles that obey quantum exclusion statistics with statistical parameters $g=1$, $g=1$, and $g=2$, respectively, subject to the constraint that the numbers of $g=1$ (fermions) exclusion particles of two types are equal.

math-ph↗

arXiv2022

Combinatorics of generalized Dyck and Motzkin paths

We relate the combinatorics of periodic generalized Dyck and Motzkin paths to the cluster coefficients of particles obeying generalized exclusion statistics, and obtain explicit expressions for the counting of paths with a fixed number of steps of each kind at each vertical coordinate. A class of generalized compositions of the integer path length emerges in the analysis.

math-ph↗

arXiv2021

Algebraic area enumeration of random walks on the honeycomb lattice

We study the enumeration of closed walks of given length and algebraic area on the honeycomb lattice. Using an irreducible operator realization of honeycomb lattice moves, we map the problem to a Hofstadter-like Hamiltonian and show that the generating function of closed walks maps to the grand partition function of a system of particles with exclusion statistics of order $g=2$ and an appropriate spectrum, along the lines of a connection previously established by two of the authors. Reinterpreting the results in terms of the standard Hofstadter spectrum calls for a mixture of $g=1$ (fermion) and $g=2$ exclusion whose physical meaning and properties require further elucidation. In this context we also obtain some unexpected Fibonacci sequences within the weights of the combinatorial factors appearing in the counting of walks.

math-ph↗