Robotics
FlashDexRetarget: Accelerating Dexterous Manipulation Data Generation through Multi-Motion Retargeting
Kyungmin Lee, Sibeen Kim, Dongyoon Hwang et al. (+6 more)
Human hand-object demonstrations provide a scalable source of data for dexterous
robot learning, but transferring them across embodiments requires physically feasible retargeting. Existing physics-based methods typically optimize each demonstration independently, leading to either limited success under finite simulation budgets or training costs that grow with
dataset size. We introduce FlashDexRetarget, an RL framework for multi-reference dexterous retargeting. We formulate retargeting as multi-reference tracking, jointly learning a single policy across many demonstrations with off-policy RL and
geometric supervision of the demonstrated interactions. This shared training formulation amortizes optimization across references while enabling the policy to track diverse hand-object interactions. On a 50-motion
benchmark from TACO, OakInk2, and HOT3D using XHand and Sharpa Wave Hand as target embodiments, FlashDexRetarget retargets 90% of demonstrations using about 30 GPU-hours, compared with about 46% at about 3,000 GPU-hours for CHORD. This corresponds to about 100 times lower training compute and a 44-percentage-point improvement in retargeting success. Ablations examine the key design choices, while experiments with up to 1,000 motions and real-world replay further demonstrate the scalability and practical applicability of our method.
Read the full abstract ▾
All authors Kyungmin Lee, Sibeen Kim, Dongyoon Hwang, Yoonsang Oh, Donghu Kim, Youngdo Lee, I Made Aswin Nahrendra, Jaegul Choo, Hojoon Lee
Quantum Physics
Trapdoored Clifford Operators and Applications
Minki Hhan and Hojune Lee
Random Clifford operators have numerous applications in
quantum computing, including randomized benchmarking, classical shadows, and
quantum authentication. However, sampling and implementing uniformly random $n$-
qubit Clifford incur near-quadratic complexity due to the size of Clifford group. We introduce a cryptographic way to overcome these barriers: trapdoored Clifford operator distributions whose samples are computationally indistinguishable from uniformly random Cliffords, yet implementing them can be much faster given the trapdoor. We construct a distribution of trapdoored Clifford operators whose elements can be sampled and implemented in near-linear time under a variant of the learning parity with noise assumption. Our constructions allow fast tableau action on Pauli labels for classical simulation, and also can be optimized to admit polylogarithmic-depth implementation. Along the way, we construct trapdoored
matrices over finite fields that support efficient multiplication by both a
matrix and its inverse, resolving an open question left by Vaikuntanathan and Zamir [SODA'26]. We use these constructions to obtain faster protocols based on random Cliffords. We also explore their applications to the worst-case to average-case reductions for
matrix and Clifford problems including the iterated
matrix multiplication and Clifford
circuit synthesis. In particular, we show the hardness of batching Clifford
circuits: synthesizing
circuits that apply the same Clifford to multiple registers is at least as hard as worst-case
matrix multiplication, even when synthesis succeeds on a small constant fraction of random Cliffords. This extends to approximate implementations by general
quantum circuits.
Read the full abstract ▾
Software Engineering
Detecting Inconsistencies in Model Specifications with LLM-as-Verifier Reasoning
Zichen Xie, Mrigank Pawagi, Lize Shao et al. (+2 more)
Model specifications define how large
language models (LLMs) should behave, guiding
alignment training, inference-time behavior, and evaluation. Yet these specifications may themselves contain defects: two individually reasonable principles may prescribe incompatible behavior when applied to the same situation, leaving no response that satisfies both. Detecting such inconsistencies is challenging. Formalizing natural-
language specifications risks losing subtle distinctions, while behavior-based testing cannot reliably distinguish specification defects from differences in model behavior. We introduce VeriSpec, the first approach to directly detect inconsistencies in model specifications by
auditing the specification text itself. Our key insight is to preserve the specification in natural
language while using an LLM as a verifier. VeriSpec extracts structured, context-aware rules, constructs a topic-guided graph to cluster behaviorally related rules at the same authority level, and applies LLM-as-verifier
reasoning to detect inconsistencies. Applying VeriSpec to the OpenAI Model Spec, we extract 405 rules and manually validate five inconsistencies, all reported to its developers, who responded positively and have initiated internal discussions. Compared with five baselines, VeriSpec identifies the most validated inconsistencies, achieves the highest precision (38.5%), and incurs the lowest cost per validated inconsistency ($11.12). These results establish direct specification
auditing as a practical complement to behavioral
alignment evaluation, catching defects at the source before they shape any model. The code is available at https://github.com/HIPREL-Group/VeriSpec.
Read the full abstract ▾
All authors Zichen Xie, Mrigank Pawagi, Lize Shao, Yang Hu, Wenxi Wang
Sound
Beyond Decodability: Do Acoustic Factors Drive Predictions in Speech-Based Alzheimer's Assessment?
Serli Kopar, Alkis Koudounas, Roshan P. Rane et al. (+3 more)
Speech-based Alzheimer's disease (AD) assessments increasingly rely on pretrained self-supervised learning (SSL) models that learn acoustic representations directly from raw audio, exposing the model to recording factors. We ask whether such factors are merely encoded in SSL representations or can systematically alter predictions. Using ADReSSo and three large SSL backbones, we apply controlled noise and reverberation interventions to participant-
speech-only, non-
speech, and full-recording audio. We combine layer-wise linear decoding, input- and representation-space interventions, and
geometric alignment analysis to distinguish acoustic decodability from influence on AD prediction. Our results show that controlled acoustic interventions alter AD predictions across all three SSL backbones. Noise, despite showing no significant diagnostic-group difference in the original data, produces the strongest intervention effects. Importantly, these effects are systematically structured relative to the classifier's decision direction, replicate on the held-out test set and reverse when the representation-space intervention direction is reversed. Together, these findings show that high predictive performance and the absence of a significant diagnostic-group difference in a measured acoustic factor are not sufficient for robustness. We argue that intervention-based robustness tests should become standard for trustworthy clinical
speech models.
Read the full abstract ▾
All authors Serli Kopar, Alkis Koudounas, Roshan P. Rane, Sam Gijsen, Paula A. Perez-Toro, Kerstin Ritter
Artificial Intelligence
Temporal-Difference Learning for Dragonchess
Jim O'Connor, Annika Hoag, Sarah Goyette et al. (+1 more)
Our research investigates how two adaptive AI methods, evolutionary transfer learning and TD(lambda), perform in the three-dimensional chess environment Dragonchess. The game challenges players with its unique board structure and computational load, making it an ideal setting to study how adaptive methods can update evaluation heuristics in novel environments. In this work we re-implement the Dragonchess engine, changing it from a PyGame engine to C++. This enables faster gameplay, allowing us to run 10,000 games with confidence intervals and significance tests, rather than a single small tournament. Both adaptive methods outperform all other
agents in the round-robin tournament. Our results showed that there is no significant difference in the performance between the evolved and learned evaluations. This research establishes the efficacy of adaptive methods in structurally complex, novel game domains.
Read the full abstract ▾
All authors Jim O'Connor, Annika Hoag, Sarah Goyette, Gary B. Parker
Functional Analysis
Ballistic Speed and Potential-Theoretic Recurrence for Homogeneous Open Quantum Random Walks
Ameur Dhahri, Chul Ki Ko, Farrukh Mukhamedov et al. (+1 more)
Our study connects root-mean-square ballistic transport to potential-theoretic recurrence for finite-range open
quantum random walks. In homogeneous walks with primitive local channels, finitely many simple periodic
Fourier peripheral
eigenvalues, no nonzero stationary
Fourier mode, and a nondegenerate quadratic spectral term, we prove a periodic uniform local limit
theorem and recover exponential finite-set return bounds for nonzero drift. The centered case yields strong Green-function asymptotics in at least three dimensions, as well as potential-kernel asymptotics in one and two dimensions. According to these assumptions, the RMS ballistic speed equals the drift norm, nonzero drift indicates transience, and centered walks are recurring in effective dimensions one and two but transitory in higher dimensions. The low-dimensional finding shows a recurrence of the origin projection in TOM. For reducible walks, an explicit harmonic \(h\)-transform converts each absorption component to an OQRW, providing a detailed breakdown of Green occupation potentials. The squared RMS speed is the absorption-weighted mean of squared component drifts, but the reduced drift-dimension classification also needs specific component return estimations. A centered noncommuting family validates the fundamental spectral assumptions in all dimensions and provides explicit potential constants. Exact finite traps and sparse reflecting barriers provide a complementary nonhomogeneous method for zero speed and TOM recurrence.
Read the full abstract ▾
All authors Ameur Dhahri, Chul Ki Ko, Farrukh Mukhamedov, Hyun Jae Yoo
Optimization and Control
Optimal Stochastic Bilevel Optimization with First-Order Oracles
Linxuan Pan and Junchi Yang
We study nonconvex–strongly-convex bilevel optimization under a
stochastic first-order oracle. We introduce MRT-FD, a single-loop first-order method that simultaneously tracks the upper-level variable, the lower-level solution, and the auxiliary response arising from implicit differentiation of the hyperobjective. MRT-FD performs one update of each variable per iteration and approximates the second-order derivative actions using order-$p$ finite differences. For any fixed finite smoothness order $p\ge1$ in the lower-level variable, MRT-FD finds an $\varepsilon$-stationary point using $\mathcal{O}(\varepsilon^{-4-2/p})$
stochastic gradient queries. We also prove a matching $\Omega(\varepsilon^{-4-2/p})$ oracle lower bound. The lower-bound construction starts from a hard nonconvex minimization chain with a stronger
stochastic oracle, and lifts it to a bilevel problem through a sinusoidal coupling with a scalar lower-level variable. Consequently, the dependence on $\varepsilon$ is optimal for every fixed finite $p$, closing the upper–lower complexity gap in this
stochastic first-order oracle setting.
Read the full abstract ▾
Artificial Intelligence
On the Divergence of Accuracy and Mechanism Consistency in Time Series World Models
Haochen Zhang, Jiaheng Guo, Zhen Xu et al. (+4 more)
A time series world model (TSWM) predicts a controlled system's state from its observed history and planned actions and exogenous inputs. Current approaches build forecasters with actions as
covariates, trained and evaluated on prediction error under the executed plan. Yet world models compare unexecuted plans, but their responses to changed plans remain untested. We ask which design choices matter and whether accurate forecasters respond to changed plans as real systems do. We address both with a formalization and
benchmark. The formalization separates state, actions and exogenous inputs, distinguishes continuous, mode and event actions, and introduces mechanism consistency, a metric built on declared action-state relations with known directions, such as a vasopressor raising blood pressure: it checks whether shifting an action moves the forecast in the declared direction. The
benchmark consolidates eight public
datasets with real actions from engineered infrastructure and clinical care, varying prediction space, plan fusion and plan encoding across seven backbones and five seeds. First, a frozen latent prediction space lowers MAE by 9.9% over observation space and gated output fusion lowers it by 12.7% over input concatenation on average, with both improving all eight
datasets; temporal plan encoding changes average MAE by at most 2.2%. Second, prediction error and mechanism consistency diverge: the lowest-error configuration is at or below chance in consistency on four of five
datasets with declared mechanisms, and no design choice avoids this. Finally, directional supervision, a loss penalizing the wrong-signed part of the response to a shifted action, significantly raises consistency on penalized mechanisms with no change in MAE. Together they give TSWMs a recipe: a frozen latent space and output-side fusion for accuracy, and a training objective for mechanism consistency.
Read the full abstract ▾
All authors Haochen Zhang, Jiaheng Guo, Zhen Xu, Zachary Plotkin, Nicholas Konz, Zhen Tan, Tianlong Chen
Mathematical Physics
Quantum hierarchies of the ILW and BO equations and the DR cycles on the moduli spaces of curves
Alexandr Buryak and Abner Hernandez Rodriguez
In an earlier paper of P. Rossi and the first author, a construction of a wide class of
quantum hierarchies was presented, with the input data given by an arbitrary cohomological field theory. A crucial role in the construction is played by the double ramification (DR) cycles on the
moduli spaces of stable curves. As a special case, one gets a quantization of the hierarchy of the Intermediate Long Wave (ILW) equation. The ILW equation has two famous limits: the Korteweg–de Vries equation and the Benjamin–Ono (BO) equation, and for the latter one a
quantum hierarchy was constructed by Nazarov and Sklyanin. In our paper, we propose a
conjecture describing a relation between a limit of the
quantum ILW hierarchy and the Nazarov–Sklyanin hierarchy and present various checks to support it.
Read the full abstract ▾
Computational Physics
RFBniCS: An open-source simulation framework for redox flow batteries
Amirhossein Aghabarari, Jørgen S. Dokken, Martin T. Horsch et al. (+2 more)
We present RFBniCS, an open-source finite-element framework implemented in FEniCSx for simulating redox-flow-battery (RFB) half-
cells. RFBniCS solves an established macro-homogeneous porous-electrode model accounting for strongly coupled electrolyte flow, multicomponent species transport, ionic and electronic charge conservation, and interfacial Faradaic charge transfer. Different from other open-access tools, RFBniCS can perform transient RFB half-
cell simulations in one-, two-, and three-dimensional geometries while explicitly resolving the transport of both redox-active and supporting-electrolyte species. It also describes electrolyte flow through porous electrodes and, in the three-dimensional formulation, through adjacent flow channels. To verify the implementation, RFBniCS's predictions in parameter-limiting regimes are compared against other published implementations, with RFBniCS generally showing superior accuracy and speed. We further demonstrate RFBniCS's capabilities by a simulation of the transient response of the negative half-
cell of a vanadium redox flow battery, with a moderately concentrated supporting electrolyte. In the current implementation, RFBniCS provides the computational basis for studying redox-flow-battery half-
cells and can be extended to complex flow-field designs, alternative chemistries, full-
cell coupling, detailed membrane transport, and additional multiphysics effects.
Read the full abstract ▾
All authors Amirhossein Aghabarari, Jørgen S. Dokken, Martin T. Horsch, Eirik Valseth, Mathijs Janssen
Quantum Physics
Toward Optimal Circuit Depth for Geometrically Local Hamiltonian Simulation
Zheyu Shen, Yusen Wu, Penghui Yao et al. (+2 more)
Understanding the dynamics of interacting
quantum systems is a central goal of
quantum simulation. For geometrically local
Hamiltonians, locality suggests that digital simulation should retain much of the parallelism of the underlying dynamics. For a target accuracy $\epsilon$, we construct a nearest-neighbor
quantum circuit that simulates the dynamics of an $n$-
qubit, time-independent, finite-range
Hamiltonian with normalized local interaction strength on a fixed-dimensional
lattice for time $T\ge1$, with
circuit depth $O(T\log(nT/\epsilon))$ and gate count $O(nT\log(nT/\epsilon))$. This reduces the polylogarithmic overhead of Haah, Hastings, Kothari, and Low to a single logarithm. On logarithmic-scale blocks from the Lieb–Robinson decomposition, a shallow but highly accurate implementation is essential for reducing the
circuit depth. Our key observation is that spatial localization also controls the cost of high precision. In each block, the error of a constant-order product formula is compensated for by a correction, yielding exponential accuracy. The correction can be implemented with depth proportional to the block diameter, introducing no additional asymptotic depth scale beyond that required by spatial localization. Furthermore, we prove an unconditional
circuit-depth lower bound for uniform XXZ
Hamiltonians: at fixed
evolution time $T$ in the stated regime and sufficiently small fixed error $\epsilon$, any nearest-neighbor
quantum circuit simulating these dynamics requires depth $\Omega(\log n/\log\log n)$. This matches our $O(\log n)$ upper bound up to a doubly logarithmic factor, establishing near-optimal
circuit depth for
lattice Hamiltonian simulation in this regime. Together, these results clarify the role of locality and precision in determining the
circuit complexity of
quantum dynamics and suggest broader principles for spatially constrained
quantum computation.
Read the full abstract ▾
All authors Zheyu Shen, Yusen Wu, Penghui Yao, Xiao Yuan, Yukun Zhang
Rings and Algebras
The Derived Algebra of Nonlinear Polynomial Divergence-Free Vector Fields
Chao Ma
We study the Lie
algebra $L_{\ge2}$ of divergence-free
polynomial vector fields on $k^n$, $n\ge3$, with coefficients of degree at least two, graded by coefficient degree. Over every field its derived
algebra in degree $d\ge3$ is the space of exact fields, those whose contraction with the volume form is an exact form, and it is already spanned by brackets with quadratic fields. In characteristic zero this is the whole degree-$d$ component. In characteristic $p>0$ the abelianization is nonzero above degree two exactly in the degrees $d\ge(p-1)(n-1)$ with $d\equiv1-n\pmod p$, and Cartier descent identifies it with a Frobenius twist of a rational $\mathrm{GL}_n$-module, tensored with a power of the determinant. For $p\ge5$ each exact component is obtained from the previous one by bracketing with quadratic fields.
Read the full abstract ▾
Number Theory
Averages of Artin characters and their relation to the Davenport–Heilbronn theorem
Victor Ahlquist
We bound the error term in the counting function of cubic extensions of $\mathbb{Q}$ by $\mathcal{O}_\epsilon(X^{\theta+\epsilon})$, with $\theta < 2/3$, improving upon results due to Bhargava, Taniguchi, and Thorne. The proof proceeds by relating a certain discriminant-reducing sieve, due to Bhargava, Shankar, and Tsimerman, to averages of Artin characters associated with the Artin $L$-function $\zeta_K(s)/\zeta(s)$, with $K$ a cubic field. Conditional on the Generalised Riemann Hypothesis (GRH) for Dedekind
zeta functions $\zeta_K(s)$, we obtain further savings and show that the error term is $\mathcal{O}(X^{2/3-1/51+\epsilon})$. Moreover, for the analogous smooth counting problem, we obtain a bound $\mathcal{O}(X^{1/2+\epsilon})$ for the error term, conditionally improving results of Shankar, Södergren, and Templier. Our arguments can also handle finitely many splitting conditions. The methods we use to study cubic fields are natural extensions of methods which are used to study the one-level density associated with the above family of Artin $L$-functions. In this direction, we improve the admissible support $[-\sigma,\sigma]$, from the previously known admissible value $\sigma < 2/5$, to $\sigma < 1$, conditional on the GRH. This also proves that at least $75\%$ of these $L$-functions are non-vanishing at the central point, conditionally improving upon results of Shankar, Södergren and Templier. Finally, one may study similar questions over a rational function field $\mathbb{F}_q(T)$, with $q$ coprime to $2$ and $3$. Here, we obtain a proportion of non-vanishing of at least $75\%$ for the analogous family of $L$-functions. Furthermore, we obtain a bound $\mathcal{O}\left(X^{1/2+\epsilon}\right)$ for the error term in the counting function of cubic function fields, improving previous results of the author. Over function fields, our results are unconditional.
Read the full abstract ▾
Cosmology and Nongalactic Astrophysics
Cosmological inference from a joint DESI DR1 full-shape power spectrum and bispectrum analysis
Caroline Guandalin, Prakhar Bansal, Pedro Carrilho et al. (+9 more)
The
galaxy bispectrum directly probes the non-linear gravitational
evolution of large-scale structure (LSS) and can break parameter degeneracies that remain in power-spectrum analyses. We present a joint full-shape cosmological analysis of three luminous red
galaxy (LRG)
redshift bins and the
quasar (QSO) sample from the first Data Release (DR1) of the Dark Energy Spectroscopic Instrument (DESI). We model the
redshift-space power spectrum at one loop and the tree-level bispectrum within the Effective Field Theory of LSS. The bispectrum is decomposed in the Tripolar Spherical Harmonics basis, for which the convolution with the survey window function can be formulated as a direct linear transformation of the theoretical multipoles. Our power-spectrum constraints are in good agreement with the official DESI DR1 full-modelling results. We investigate the impact of including the bispectrum in the inference, finding that the monopole substantially improves the constraints on the cold dark matter density and amplitude of matter fluctuations by 9-18% and 8-20%, respectively, in the individual-tracer analyses. The corresponding reductions are 15% and 10% for the combined LRG sample, and 6% and 4% when all tracers are combined. In a restricted test using the first LRG bin, the bispectrum quadrupole changes the marginalised uncertainties by only a few percent. Extending the analysis to $w_0w_a$CDM substantially broadens the cosmological posteriors, while the bispectrum produces only a mild change in the allowed dark-energy parameter region, which remains sensitive to the adopted prior ranges. Our results demonstrate the potential of higher-order clustering statistics to improve cosmological constraints, while providing a framework for incorporating the bispectrum into full-shape analyses of current and future spectroscopic
galaxy surveys.
Read the full abstract ▾
All authors Caroline Guandalin, Prakhar Bansal, Pedro Carrilho, Alejandro Aviles, Mike Shengbo Wang, Marcos Pellejero-Ibañez, Aaditya Sarma, Jaide Swanson, Marco Bonici, Florian Beutler, Arnaud de Mattia, Hee-Jong Seo
Atmospheric and Oceanic Physics
Varda-single-1.0: deterministic data-driven weather forecasting at 1 km resolution over Switzerland's complex topography
Alberto Pennino, Francesco Zanetta, Michele Cattaneo et al. (+28 more)
We present Varda-single-1.0, a medium-range data-driven weather prediction system built for the Alpine domain. It provides hourly deterministic regional forecasts on a mesh of 1 km resolution and global forecasts on a 31 km mesh. The system comprises two independently trained stretched-grid Graph
Transformer models with encoder-processor-decoder architecture, developed in the Anemoi framework: a 6-hourly autoregressive forecaster and a temporal downscaler reconstructing hourly forecasts between the forecaster's steps. Its training curriculum includes pre-training on ERA5 reanalysis data, followed by training on a 20-year kilometre-scale regional reanalysis, and finally fine-tuning on operational kilometre-scale analyses. Verified over one year against operational analyses and surface station observations, Varda-single is competitive with or improves on MeteoSwiss' operational numerical weather prediction baselines for most headline scores and variables. It broadly matches the skill of the high-resolution 1 km ICON-CH1-EPS control at lead times up to +33 h and generally outperforms the 2 km ICON-CH2-EPS control at lead times up to +120 h. Despite competitive aggregate scores, Varda-single underestimates some local wind maxima and produces overly smooth convective precipitation fields, consistent with the smoothing associated with squared-error training. To gain insight into the model's behaviour, we investigate three case studies beyond the aggregated headline scores, and find particular weaknesses in Varda-single's representation of local winds over complex terrain. Varda-single represents an important step in the development of high-resolution ML
forecasting over complex terrain, in complementing the operational regional numerical weather prediction models of MeteoSwiss with data-driven models and in providing a pretrained model for researchers and user-specific applications.
Read the full abstract ▾
All authors Alberto Pennino, Francesco Zanetta, Michele Cattaneo, Claire Merker, Radi Radev, Jonas Bhend, Louis Frey, Hugues de Laroussilhe, Ophélia Miralles, Carlos Osuna, Daniele Nerini, Andreas Pauling, Daniel Hupp, Ulrich Hamann, Mary McGlohon, Marti Bosch, Luca Lanzilao, Marco Arpagaus, Lukas Jansing, Daniel Leuenberger, Mark A. Liniger, Katrin Ehlert, Matthew Chantry, Håvard Homleid Haugen, Gert Mertes, Ana Prieto Nemesio, Mario Santa Cruz, Jasper Wijnands, Gabriel Moldovan, Harrison Cook, Oliver Fuhrer
Artificial Intelligence
Code Owns the Simulation, Jev Owns the Evaluation
Yaodong Yang, Hongyao Tang, Yi Ma et al. (+4 more)
Judgment models such as \jev{} return, in a single call and without
reasoning text, a probability for each described option. This makes them attractive as an
agent's action-selection layer, but it is unclear which decisions they can be trusted with. We test \jev{} on reflection tests, one-shot
matrix games, the text game ALFWorld and
robot control, and find a sharp boundary. \jev{} succeeds when the right option can be judged from what the input describes, which we call evaluation. Specifically, it solves 99% of the counterintuitive Cognitive Reflection Test questions. However, it fails when the right option depends on simulation (i.e., predicting something not in the input), such as the opponent's action or the subgoal that must come first. In games, \jev{} plays suboptimally as if its rational opponent acted at random, because the opponent's action is not given. In ALFWorld, \jev{} favors commands that mention an object or place named in the task description. For example, given the task “put a clean knife in the drawer”, \jev{} carries an unwashed knife straight to the drawer instead of first washing it at the sink. Surprisingly, many of these failures are not due to a lack of knowledge. Asked separately what the opponent will do, \jev{} usually answers correctly, and it responds well given the opponent's action. It fails when one call must both perform the simulation and evaluate based on it. This suggests letting code make the prediction or simulation. When code supplies it, such as a lookahead in ALFWorld and physics simulation in
robot control, \jev{} becomes an expert controller through its general evaluation ability.
Read the full abstract ▾
All authors Yaodong Yang, Hongyao Tang, Yi Ma, Xingyu Fan, Weixun Wang, Jinpeng Li, Tianpei Yang
Artificial Intelligence
Continuous Process-Level Evaluation for Evolving Enterprise AI Agent Skills
Ngoc Phuoc An Vo, Aarya Doshi and Vadim Sheinin
Enterprise AI
agent skills evolve as tool APIs, models, and specifications change, yet final-output evaluation can miss process-level behavioral drift. We present a continuous evaluation framework combining outcome-level and process-level checks, applied to Revenue and Productivity variants of a Business Value Determination skill in an enterprise Value Aware Resiliency system. The framework independently computes per-run ground truth, materializes reusable template tests, and evaluates tool selection, arguments, execution order, and database integrity through programmatic checks and a narrowly scoped LLM judge. We evaluate 240 trials across two skills, two specification variants, two
agent harnesses, and three models. Of 175 trials passing all applicable final numerical checks, 162 (92.6 percent; Wilson 95 percent CI: 87.7-95.6 percent) contained another evaluator-detected deviation. Under a broader seven-check final-state definition, 151 of 164 passing runs (92.1 percent; 95 percent CI: 86.9-95.3 percent) still violated a trajectory check. Dependency attribution reduced a mean of 6.34 failed checks per run to 2.65 roots. Specification sensitivity varied by model and harness, with exploratory bootstrap interaction intervals excluding zero for all three Revenue comparisons and one of three Productivity comparisons. Runtime-resolved templates provided reusable regression coverage across the evaluated configurations; longitudinal validation under actual API
evolution remains future work.
Read the full abstract ▾
Quantum Physics
Experimental realization of Wheeler's delayed-choice experiment with dual selections
Xiaowan Yang, Xinglei Yu, Liangsheng Li et al. (+5 more)
Wheeler's delayed-choice experiment demonstrates wave-particle duality of particles using different experimental configurations of a Mach-Zehnder interferometer. In a
quantum version of this experiment, the wave-particle behavior of
photons can be observed by controlling presence or absence of the second beam splitter. Here, we implement a delayed-choice experiment with dual selections based on entangled
photons, experimentally controlling both of two beam splitters in the Mach-Zehnder interferometer, where the presence and absence of the second beam splitter are simultaneously controlled through path encoding. Our experiment reveals the wave-particle behavior of single
photons under different configurations of two beam splitters. We further discuss the scenario when two beam splitters are in a
quantum superposition of being present and absent,
photons will be in a wave-particle
quantum superposition state.
Read the full abstract ▾
All authors Xiaowan Yang, Xinglei Yu, Liangsheng Li, Xinzhi Zhao, Tianle Zheng, Chengjie Zhang, Chuan-Feng Li, Guang-Can Guo
Machine Learning
Pooling Helps, Learned Weighting Hurts In-Context: Decomposing Group Attention
Michael Fore, James Mason Inder, Mrishika Nair et al. (+2 more)
Group
attention, introduced by the time series
forecasting model Chronos-2, attends over the variates of a group at a fixed patch index and serves both multivariate (MV) and in-context learning (ICL)
forecasting. Rather than evaluating this cross-variate
attention design as a whole, we ask which part of the mechanism earns the benefit and probe its applicability to both MV and ICL regimes. By editing the
attention matrix $\alpha$ at inference we separate the two pathways a head comprises: V/O, which projects a weighted summary of the group, and Q/K, which decides the weights. Uniform pooling (V/O without any Q/K weighting) is positive on 18 of our 20 sensor-network configurations, while the learned weighting (Q/K) splits by group type: its contribution is positive or negligible for MV, but materially degrades 8 of the 10 sensor-network ICL configurations, leaving 4 of them worse than univariate inference. By isolating the impact of different layers, we find that uniforming $\alpha$ in the first block alone improves every ICL configuration we test.
Read the full abstract ▾
All authors Michael Fore, James Mason Inder, Mrishika Nair, Praneetha Vaddamanu, Sharlina Keshava
Mesoscale and Nanoscale Physics
Flux- and tunnel-parity-controlled Josephson nonreciprocity from singlet-triplet competition in a parallel double quantum dot
Debika Debnath, Fernando Dominguez and Patrik Recher
Nonreciprocal supercurrents in interacting nanostructures can arise when the positive and negative critical currents probe inequivalent portions of a phase-dependent many-body spectrum. We study this mechanism in a parallel double-
quantum-dot Josephson junction with strong intradot Coulomb repulsion, where local and non-local Cooper-pair transfer coexist. Using exact diagonalization of a zero-bandwidth
Hamiltonian with explicit superconducting orbitals, we show that orbital flux and the gauge-invariant tunnel-sign parity control the interference between local Josephson processes and non-local exchange. For detuned dots, this produces flux-tunable singlet, triplet, and doublet ground-state spectral branches and two rectification regimes. Near the boundaries of the charge sector with singly occupied dots [(1,1) sector], doublet branches compete with singlet or triplet branches, yielding diode efficiencies approaching 40%. Within the triplet-dominated (1,1) sector, nearby singlet crossings asymmetrically reshape the ground-state envelope even when both critical-current extrema lie on the triplet-like branch, producing a broader response of order 10-15%. Finite temperature generally suppresses charge-crossover rectification, however, for intermediate temperatures we find an enhancement of both the normalized efficiency and the absolute critical-current asymmetry in the triplet regime. These results establish phase-dependent many-body branch competition as a gate-, flux-, and tunnel-parity-sensitive source of Josephson nonreciprocity.
Read the full abstract ▾
Algebraic Geometry
Cohomology of the fine compactified universal Jacobians $\overline{J}_2$ and $\overline{J}_{2,1}$
Marco Fava
We compute the rational
cohomology of the fine compactified universal Jacobians over the
moduli spaces of stable curves of genus $2$ with $0$ and $1$ marked points.
Computation and Language
Haricharan Balasundaram and V. Arvind Rameshwar
Language model (LM)
alignment broadly aims to perturb a given LM $Q$ into an aligned LM $q$ such that i) the outputs produced by $q$ and $Q$ are 'close' in probability, ii) $q$ has a higher expected reward than $Q$. Two common techniques for LM
alignment are: KL-constrained RL, which requires knowledge of the LM distribution and is computationally expensive, and the best-of-$n$
algorithm, which requires only sampling from the LM. The work of Yang et al. established asymptotic closeness between the distributions produced by the two
alignment methods for an $m$–length i.i.d.
token sequence output by the LM, in the limit as $m$ increases to infinity. However, the i.i.d. assumption is not representative of practical LMs, whose output sequences often have
memory. In this paper, we extend the asymptotic closeness result to the case when the $m$–length
token sequence outputted by the LM is Markovian. Further, for finite-length output sequences – particularly, when $m=1$ – we provide a complete characterization of LM distributions and reward functions for which the KL-divergence between the distributions produced by the two
alignment methods is zero – a question first posed in Yang et al.
Read the full abstract ▾
Machine Learning
Scientific Discovery under Validation Congestion via Multi-Fidelity Pairwise Rankings
Kevin Tirta Wijaya, Alston Lo, Michael Sun et al. (+2 more)
Modern computational methods can now propose candidate molecules, materials, and other scientific designs at an unprecedented scale, creating a validation congestion where candidates are abundant, but experimental capacity to physically evaluate them remains scarce. Discovering novel scientific designs has therefore become increasingly dependent on curation: selecting a small set of promising designs for slow and costly experiments. Existing curation methods typically rely on data-driven regression models that predict absolute scores, but training these models requires substantial experimental data to begin with. Yet, useful curation signals do not have to take the form of absolute measurements, as scientific design discovery is often comparative in nature. Here, we propose that curation can instead be primarily driven by expert pairwise rankings, which are substantially easier to gather. The expertise can come from computational tools or human input of multiple levels of fidelity, ranging from empirical rules of thumb to agentic workflows and experienced scientists. We introduce PRISMS, a framework that uses pairwise rankings from one or more experts, potentially spanning multiple levels of expertise, to identify the most promising candidates without relying on data-hungry regressors. When experts differ in fidelity and cost, PRISMS escalates pairwise queries from lower- to higher-fidelity rankers based on a Fisher-information criterion. In iterative screening that selects designs from fixed drug discovery libraries, PRISMS achieves 50% top-10 discovery recall in ~42% fewer rounds than regression-only active learning, and in ~15% fewer rounds than the ranking-based method with no selective escalation. In optimization that generates new designs without restriction to a predefined library, PRISMS achieves ~18.8% higher hypervolume than the
Bayesian optimization baseline.
Read the full abstract ▾
All authors Kevin Tirta Wijaya, Alston Lo, Michael Sun, Wojciech Matusik, Vahid Babaei
Signal Processing
Token Communication-Assisted Collaborative Embodied Artificial Intelligence: Concepts, Framework, and Opportunities
Peng Yi and Ying-Chang Liang
Collaborative
embodied artificial intelligence (CEAI) enables multiple physical
agents to perceive, reason, and act cooperatively in dynamic environments. Effective communication is essential for CEAI, yet CEAI
agents must exchange not only large
multimodal observations but also task-relevant insights, intents, and interactive information over long horizons. This article investigates
token communication (TokCom) as a native intelligence interface for CEAI, in which
tokens serve jointly as compact semantic carriers for communication and fundamental inference units for generative foundation models (GFMs). We first discuss how TokCom supports insight sharing, intent
alignment, and interactive control among
embodied agents. We then propose a TokCom-assisted CEAI framework driven by a task-adaptive communication protocol. Comprising a compact codebook, syntax rules, and contextual examples, this protocol guides GFM-based transceivers to distill messages into compact
tokens and reconstruct them after wireless transmission. A case study on collaborative object transport demonstrates that the proposed TokCom framework substantially reduces the source payload bit consumption while preserving task efficiency and showing robustness under noisy channels. Finally, we outline future research directions.
Read the full abstract ▾
High Energy Physics (Phenomenology)
Near Threshold Kaluza-Klein Graviton Decays and the Dark Dimension
Kevin Langhoff
Near threshold, intra-tower decays of Kaluza-Klein (KK) gravitons are pure $d$-wave, with rate $\Gamma \propto p^5$ in the decay products momentum $p$, rather than the $s$-wave $\Gamma \propto p$ assumed in the dark dimension literature. In a flat dark dimension this excludes KK gravitons as a dark matter candidate. Even if some other source of dark matter is assumed, the irreducible freeze-in abundance of KK gravitons produce MeV gamma rays which allows one to set upper limits on the reheating temperature; additionally using the lower bound on the reheating temperature from big-bang nucleosynthesis and the
cosmic microwave background requires the radius of the extra dimension to be almost three orders of magnitude smaller than the distance
conjecture would naively prefer, $R\sim \Lambda^{-1/4}$.
Read the full abstract ▾
Quantum Physics
QPI-DeepONet-MAC: A Scalable and Stable Hybrid Classical-Quantum Architecture for Physics-Informed Deep Operator Networks
Said Lantigua, José Valencia, Gilson Giraldi et al. (+2 more)
General operator learning for parametric partial differential equations (PDEs) is a fundamental challenge at the intersection of artificial intelligence and physics-based modeling. Physics-informed Deep Operator Networks (PI-DeepONets) incorporate governing equations into learning, but face optimization difficulties in high dimensions.
Quantum extensions offer additional representational capacity, yet their trainability can be hindered by barren plateaus. Here, we introduce QPI-DeepONet-MAC, a hybrid
quantum-classical architecture that integrates parameterized
quantum circuits into PI-DeepONets through a multiplicative-and-additive coupling (MAC) of classical representations and
quantum expectation values. We prove analytically that the proposed architecture retains universal operator approximation capabilities. Furthermore, we derive bounds on
quantum parameter gradients, establishing explicit conditions to avoid barren plateaus alongside a scaling criterion for trainability. Additionally, we derive a lower bound on the number of sensors required for a target accuracy, directly linking sampling requirements to input function regularity, domain dimension, and PDE dynamical growth. Finally, we propose an informational regularization scheme based on
quantum coherence and state fidelity dynamics to prevent coherence loss and preserve
quantum structure during training. Together, these results establish a theoretical foundation for
quantum-enhanced physics-informed operator learning and position QPI-DeepONet-MAC as a scalable, stable framework for solving parametric PDEs.
Read the full abstract ▾
All authors Said Lantigua, José Valencia, Gilson Giraldi, Renato Portugal, Jonas Maziero
Statistical Methodology
Generalized Engression Models
Xinwei Shen, Zijian Guo and Francis Bach
We consider estimating the conditional distribution of a multivariate outcome given
covariates when its coordinates may be continuous, binary, categorical, ordinal or rankings, and are conditionally dependent on one another. Different statistical methods have been developed for each outcome type, and most of them target a summary of the conditional distribution, such as the mean of each coordinate, rather than the joint distribution of the outcome vector. We develop generalized engression models, a unified nonparametric distributional regression framework for outcomes of any type. The proposed method builds upon engression, a scoring-rule-based deep generative model, and introduces a data-type-specific link function and a
stochastic perturbation that smooths the loss, enabling gradient-based training even with discontinuous links. We establish universal representation results for continuous, discrete and mixed outcomes. In simulations and in two applications, 242 species in a community ecology
benchmark and a 17-dimensional mixed-type health outcome, the method matches type-specific models on marginal scores, improves on them on the joint distribution, and matches or exceeds purpose-built state-of-the-art joint species distribution models. Software is available in Python.
Read the full abstract ▾
Systems and Control
Differentiable Hybrid-Action Neural Feedback Control for District Heating Networks
Nicolas Kirsch, Corrado Sgadari, Alessio La Bella et al. (+1 more)
Many cyber-physical systems require control policies that combine continuous setpoints with discrete operational de- cisions, such as equipment switching, mode selection, or resource scheduling. Discrete actions are not differentiable, which ob- structs gradient-based policy training, while conventional mixed- integer formulations remain costly to solve online. This paper proposes a hybrid-action neural controller (HANC) in which a continuous branch, a categorical branch and a differentiable assembly layer jointly generate commands that satisfy complex actuator constraints by construction. Categorical decisions are handled using a straight-through Gumbel
estimator, enabling the policy to be trained by backpropagation through time over full closed-loop rollouts. The proposed framework is deployed on a district heating network (DHN) featuring multiple heat generation units and stratified thermal energy storage. Its performance is evaluated on a simulation of a real DHN located at RSE SpA in Italy. The resulting policy jointly learns switching decisions and continu- ous operating setpoints. Under dynamic electricity pricing, the learned controller reduces operating cost by 30% compared to a rule-based industrial baseline. We also show that, compared with a deterministic straight-through relaxation, injecting noise during training achieves similar cost while reducing hard switching by an order of magnitude, and attribute this difference to the wider decision margins of the resulting policy.
Read the full abstract ▾
All authors Nicolas Kirsch, Corrado Sgadari, Alessio La Bella, Giancarlo Ferrari-Trecate
Machine Learning
Beyond Linear Concepts: Discovering and Aligning Non-Linear Concept Manifolds in Large Language Models
Tido Specht, Elias Benedict Krey, Nils Neukirch et al. (+1 more)
Understanding information processing in large
language models (LLMs) requires dissecting the
geometric organization of their internal
token representations. While existing mechanistic interpretability (MI) methods seek to extract concepts, they are constrained by a strong linearity assumption challenged by evidence of non-linear feature
manifolds. We move beyond linear concepts by adapting Non-Linear Multi-Dimensional Concept Discovery (NLMCD) from computer
vision to
token-level LLM activations, modeling concepts as low-dimensional
manifolds. To compare concept
manifolds across layers and models, we introduce a concept-based
alignment (CBA) score, a generalized Rand index that measures
geometric proximity without explicit feature matching. Our analysis yields six key findings: (i) a neighboring-layer sanity check shows CBA is more sensitive than PCA- or CKA-based linear baselines; (ii) layer-by-layer
alignment matrices reveal two block structures in intermediate and late layers, consistent across models and obscured by linear metrics; (iii) concept composition remains syntax-dominated through most of the network before giving way to increasingly mixed syntactic-semantic concepts in later layers, with increasing output-orientation toward the final layers; (iv)
multilingual concept sharing between English and Mandarin is training-dependent rather than universal, strongest in Qwen, weaker in Llama, and absent in GPT-2; (v) inter-model
alignment mirrors this structure, with strong correspondence between same-family Qwen models of different scale but weak
alignment across model families; and (vi) across Tulu-3 training stages,
alignment is highest between adjacent stages, with the largest shift between the base model and SFT, while subsequent preference-
alignment stages (DPO, RLVR) leave early layers largely unchanged and RLVR mostly preserves DPO's concepts in late layers.
Read the full abstract ▾
All authors Tido Specht, Elias Benedict Krey, Nils Neukirch, Nils Strodthoff
Systems and Control
Finite-Data Safety Informativity Under Dynamic Asymmetric Actuation
Abhinav Sinha, Praveen Kumar Ranjan and Yongcan Cao
When the system model is not fully known, measurement error and limited excitation can leave several models consistent with the same finite data. A command judged safe for one model may fail for another, while limited control authority can prevent the corrective action needed to preserve safety. To ensure safety under model uncertainty and asymmetric input limits, we develop a finite-data certificate that determines whether a command can enforce a prescribed safety inequality. For a linearly parameterized safety channel with exactly known regressors and bounded aggregate residual error, we derive a support formula for the worst-case safety contribution of all data-consistent models. The formula identifies the regressor directions that admit a finite bound, allowing rank-deficient records to contribute to safety certification. Using certified componentwise bounds on actuator tracking error yields an affine inequality with a necessary and sufficient test for pointwise command feasibility. The affine inequality reduces computation of the closest certified command to a scalar root-finding problem. It also yields a closed-form gate that selects the largest certified fraction of a prescribed command segment. The proposed certificate guarantees output safety within its operating domain, provided the feedback is locally Lipschitz and the uncertainty bounds remain valid. Domain retention and full-state continuation extend this guarantee to all time. A vehicle study demonstrates that output safety can be certified from finite measurements in a safety-critical setting with model and actuator uncertainty.
Read the full abstract ▾