Robotics
LLA-MPPI: Rapidly Adaptive Whole-body Control of Legged Robots with GPU-Accelerated Parallel Simulations
Sebin Jung, Maitham F. AL-Sunni, Juan Alvarez-Padilla et al. (+3 more)
Real-time whole-body controllers for legged
robots typically plan through a fixed nominal model and degrade when the deployed dynamics change. Adaptive methods typically require a model structure that contact dynamics do not provide, or they need offline training for each anticipated condition. We present Look-back and Look-ahead Adaptive Model Predictive Path Integral control (LLA-MPPI). The method converts whole-body adaptation into selection over a bank of GPU-batched contact simulators with different physical or structural parameters. Windowed prediction errors select the simulator that best explains recent motion. A whole-body MPPI planner optimizes controls through the selected model. The framework requires no offline training, and its selected hypotheses are physically interpretable. Across four simulated tasks, it achieves 97.5% success while the strongest baseline reaches 74% and an oracle with the true model reaches 98.5%. Hardware validation on a Unitree Go2 shows the
robot walking under a payload added mid-run, walking after one leg is disabled, and pushing a box to its goal while increasing its mass on the fly. Code,
videos, and project details are available at: https://lla-control.github.io
Read the full abstract ▾
All authors Sebin Jung, Maitham F. AL-Sunni, Juan Alvarez-Padilla, Zachary Manchester, Changliu Liu, John M. Dolan
Computational Physics
Self-Shielded Multigroup Microscopic Cross Section Generation Using Targeted Artificial Neural Networks
Joshua Nichols, Andrew Osborne, Ben Whewell et al. (+1 more)
Multigroup
neutron transport models are used widely in simulation codes due to their
memory and compute efficiency compared to continuous energy methods. However, multigroup codes rely on
nuclear data prepared using approximate methods that require significant manual effort and expertise to account for self-shielding effects. In addition, multigroup
nuclear data libraries prepared in this way can be limited in their applicability for modeling diverse reactors and operating conditions. Here we show that trained artificial neural networks can predict self-shielded microscopic cross sections accurately in pincell simulations containing uranium dioxide fuel with a fixed geometry and temperature. The model's predictive accuracy remains high across fuel enrichment and burnup ranges encountered in light water reactors. Critically, the neural networks' predictions require only the atomic concentration of the fuel's constituent nuclides as inputs. The networks are trained using 8,704 samples of training and validation data generated using continuous energy OpenMC pincell simulations. The trained neural networks predict self-shielded total, fission, absorption, elastic, and total scattering cross sections for combinations of 90 nuclides in the CASMO-8 energy group structure. The neural networks' predictions of self-shielded cross sections are validated by comparison with 1,500 samples of reference cross sections unseen during training and tallied using OpenMC simulations. The predicted and ground truth cross sections in the test data set agree to within 0.119% on average. When the predicted cross sections are used in a multigroup
eigenvalue run, they achieve a mean absolute error in keff of 200 pcm.
Read the full abstract ▾
All authors Joshua Nichols, Andrew Osborne, Ben Whewell, Nathan Gibson
Human-Computer Interaction
How assigned AI use before class shapes active student engagement in class
Dan J. Wang, Neelam Modi Jain, Vanessa Burbano et al. (+5 more)
AI learning tools are rapidly entering classrooms, but evidence about whether they help students learn is mixed and rests mostly on test scores. Comparatively less research addresses whether the use of AI changes students' live learning behaviors in class. Here, we report the results of a preregistered field experiment with 759 MBA students enrolled in ten sections of a course, in which each student was randomly assigned two of ten class sessions to prepare for with a purpose-built voice-based AI discussion partner. After two uses of the AI discussion partner, students made about 31% more voluntary contributions in each later class session. Students who used the AI discussion partner more also reported greater comfort speaking up and greater perceived learning, but not greater focus or motivation. These findings suggest that repeated practice with a voice-based AI partner can meaningfully increase students' engagement in class discussion, enhancing a critical intermediate learning outcome.
Read the full abstract ▾
All authors Dan J. Wang, Neelam Modi Jain, Vanessa Burbano, Jorge Guzman, Daniel Keum, Soomi Kim, Bruce Kogut, Nataliya Wright
Robotics
FoldBack: Self-Correcting Masked Generative Policy for Long-Horizon Garment Folding
Lipeng Zhuang, Shiyu Fan, Yingdong Ru et al. (+4 more)
We present FoldBack, a self-correcting masked generative policy for long-horizon garment folding. Existing long-trajectory policies may continue after a missed or slipped grasp even when the garment has not reached the intended configuration. We structure FoldBack's recovery mechanisms around three inference-time decisions: when to refine and verify, how to roll back, and where and how to retry. FoldBack aligns refinement and grasp verification with pick-and-place events, returns the
robot to a retryable pre-grasp configuration while preserving successful grasps, and selectively regenerates the failed segment and selected future actions while avoiding previous failed grasp locations. To our knowledge, FoldBack is the first editable full-trajectory policy to unify these decisions, enabling failed interactions to be detected, undone, and repaired before execution continues, without recovery demonstrations or base-policy retraining. Across 33 real garments from six categories, FoldBack achieves 75.2% final folding success and 0.837 final-mask IoU, versus 45.7% and 0.689 for the strongest prior baseline.
Read the full abstract ▾
All authors Lipeng Zhuang, Shiyu Fan, Yingdong Ru, Zhuo He, Florent P. Audonnet, Paul Henderson, Gerardo Aragon Camarasa
Differential Geometry
Sharp dynamical isoperimetric principle for mean curvature flow
Beomjun Choi, Wenkui Du, Seung Chul Park et al. (+1 more)
In this paper, we establish a sharp dynamical isoperimetric principle for weak mean
curvature flow. We prove that, among weak mean
curvature flows with fixed area of initial smooth closed hypersurfaces, the standard smoothly shrinking spherical mean
curvature flow uniquely maximizes the extinction time. More precisely, for the level set flow $K_t$ starting from the boundary of a smooth bounded domain $\Omega\subset\mathbb R^{n+1}$, as well as for the integral Brakke flow $\{\mu_t\}$ with initial Radon measure $\mu_0=\mathcal H^n\llcorner\partial\Omega$, we establish the corresponding optimal extinction time estimates \begin{equation*} T_ext(\Omega),\, T^B_ext(\Omega) \leq \frac{1}{2n} \left(\frac{P( \Omega)}{|{\mathbb{S}^n}|}\right)^{\frac{2}{n}}, \end{equation*} where $P(\Omega)$ is the perimeter of $\Omega$ representing the area of $\partial \Omega$, and the equality holds if and only if $\Omega$ is a round ball and the flow is the standard multiplicity-one smoothly self-shrinking round sphere. In particular, we obtain the sharp $L^p$-estimates for the arrival time function of a smooth bounded mean convex domain. In addition, we also establish the sharp isoperimetric inequality for the
parabolic measure of the space-time track filling $X$ of outward minimizing level set flow starting from the boundary of a smooth bounded domain $\Omega\subset \mathbb R^{n+1}$: \begin{equation*} \mathcal H_{\mathrm{par}}^{n+2}(X) \leq \frac{\pi} {2n(n+2)^2|{\mathbb{S}^n}|^{{\frac{2}{n}}}} P(\Omega)^{\frac{n+2}{n}}, \end{equation*} where the equality holds if and only if $\Omega$ is a round ball and the flow is standard smoothly shrinking round sphere.
Read the full abstract ▾
All authors Beomjun Choi, Wenkui Du, Seung Chul Park, Junseo Youn
Machine Learning
Composing What Each Teacher Learned: Delta Distillation for Multi-Teacher On-Policy Learning
Yuanda Xu, Hejian Sang, Zhengze Zhou et al. (+4 more)
Multi-teacher on-policy
distillation (MOPD) is used in two settings. In common-domain composition, several teachers score each student rollout from one
prompt domain and their signals form a single target; in routed-domain
distillation,
prompts from different domains are assigned to the corresponding specialist. Both settings usually transfer each teacher's endpoint policy, which mixes what post-training changed with preferences inherited from the teacher's base. We introduce $\Delta$-MOPD, which transfers each teacher's teacher-minus-base logit shift re-anchored at the student's frozen initialization, and compare it with endpoint supervision in both settings while holding teacher selection fixed. We first expose the mechanism that impedes endpoint transfer: inherited base pull can exceed the post-training shift. Removing it reduces the teacher-term norm ratio and target–student KL. Across our experiments, the results suggest that shift targets are particularly useful when teacher signals are combined at a state. With three composed teachers, $\Delta$-MOPD exceeds endpoint composition by $4.11$ Math and $1.95$ five-
benchmark points; with two, it matches endpoint accuracy. Under phased routing, it achieves higher mean performance in both phase orders and reduces the observed order gap from $10.50$ to $6.42$ points. Under interleaved routing, where each update involves one teacher, the two targets perform comparably. The phased results provide supporting evidence that the benefit may extend to signals accumulated across training phases. Target construction is thus an independent design axis in MOPD, complementary to teacher selection. The experiment repo could be find at https://github.com/HJSang/Delta-MOPD
Read the full abstract ▾
All authors Yuanda Xu, Hejian Sang, Zhengze Zhou, Shayan Mohajer Hamidi, Xiaomin Li, Rohit Jain, Alborz Geramifard
Machine Learning
NeuralBES: A Differentiable, Control-Aware Emulator for Scalable Building Energy Modeling
Ting-Yu Dai, Takuya Kurihana, Wing Yee Au et al. (+1 more)
Demand-side flexibility i.e.
forecasting, shifting, and curtailing residential energy loads, depends on thermal models trusted across millions of heterogeneous buildings. Existing tools force a hard tradeoff: high-fidelity physics simulators such as EnergyPlus are accurate but sequential and require per-building calibration, while purely data-driven sequence models scale but abandon the physical structure that makes their predictions trustworthy. We introduce NeuralBES (Building Energy Simulation), a differentiable emulator that resolves this tradeoff by parameterizing a resistance–capacitance (RC) based thermal model with a shared neural encoder: static building metadata such as floor area, vintage, and HVAC type is mapped to physically bounded capacitances, conductances, and equipment coefficients, which become the coefficients of a scalar linear recurrence solved via a log-space parallel scan, and a predictor–corrector loop closes the thermostat–temperature nonlinearity while preserving full-horizon gradient flow. Trained on the ResStock
dataset across three
climate zones, NeuralBES handles heterogeneous building archetypes, vintages, and
climate zones within a single trained encoder, while black-box baselines produce statistically plausible but physically inconsistent trajectories. On the annual full-year rollout, NeuralBES is the only data-conditioned model that is simultaneously physics-valid and accurate to within 4 MAPE points of the strongest raw-error baseline, while operating at roughly an order of magnitude fewer parameters than the
transformer and recurrent baselines; among physics-valid baselines at parameter parity it more than halves the MAPE of the grey-box RC alternative.
Read the full abstract ▾
All authors Ting-Yu Dai, Takuya Kurihana, Wing Yee Au, Hon Yung Wong
Analysis of PDEs
Asymptotic Stability of Domain Walls for One Dimensional Schrödinger Map with Dzyaloshinskii-Moriya Interaction
Ze Li, Zifan Wang and Lifeng Zhao
In this paper, we study the long time dynamics of small perturbations of static domain walls in weighted Sobolev spaces for the one dimensional Schrödinger map with constant Dzyaloshinskii-Moriya interaction. The perturbation equation contains derivative nonlinearities, and its linearized operator has a threshold resonance. The results are twofold. (i) We prove the asymptotic stability of the domain wall modulo the translation and rotation
symmetries. More precisely, the corresponding modulation parameters converge as time tends to infinity, while the radiation decays at the sharp rate $t^{-1/2}$. (ii) We prove that the radiation exhibits modified scattering, with an explicit logarithmic phase correction generated by the long range cubic interaction. The proof develops
geometric reduction to overcome the loss of derivatives and exploits structural cancellations in the nonlinear interactions.
Read the full abstract ▾
Computer Vision
MORCA: Offline-to-Online Reinforcement Learning for Adaptive Cache Reuse in Video Diffusion Acceleration
Yuxiang Xiong, Ruiyan Wang, Wenqiang Wang et al. (+5 more)
Diffusion Transformers (DiTs) achieve remarkable performance in
video synthesis, but their iterative denoising process suffers from high inference latency. To address this, caching has emerged as an effective acceleration strategy by capitalizing on inter-step redundancy during denoising. Existing dynamic caching methods typically estimate the error that cache reuse would introduce at each denoising step (step error) to guide cache decisions, whereas our concern is how much quality loss cache reuse would cause in the final generated
video (terminal error). We show that step error does not directly correspond to terminal error and that latent information helps capture their relationship, thereby informing cache decisions. Moreover, existing threshold-based methods cannot provide precise speedup control, making it difficult to meet practical requirements for user-specified acceleration targets. To address these limitations, we introduce MORCA, a cache scheduling framework trained through offline-to-online reinforcement learning to make latent-aware reuse/recompute decisions under user-specified acceleration targets. Extensive experiments on different
video generation models across multiple target acceleration ratios demonstrate that MORCA achieves better generation fidelity than state-of-the-art caching methods under comparable computational budgets. Code is available at https://github.com/x10ngyx/MORCA.
Read the full abstract ▾
All authors Yuxiang Xiong, Ruiyan Wang, Wenqiang Wang, Teng Hu, Songhang Shen, Bohao Feng, Hongqian Deng, Ran Yi
Theoretical Economics
Risk-minimizing GUE-implementations of truthful reporting in a Dynamic Social Choice Problem
Endre Csóka and Máté Viczián
The Guaranteed Utility Mechanism (GUM) was introduced to implement efficiency in Guaranteed Utility
Equilibrium (GUE) in dynamic social choice problems. We show that this mechanism also has a previously unknown connection to risk minimization. Fixing an efficient decision policy, we study variance-minimizing budget-balanced transfer rules in the class of GUE-implementations of truthful reporting. A symmetrized version of GUM (sym-GUM) minimizes the sum of the
agents' utility variances under truthful reporting when there is no public state. When a public state is present, a slight modification of this rule yields the minimizer for the same objective. Sym-GUM, however, does not minimize transfer risk. The minimizer for that problem is instead the symmetrized version of a closely related GUM-like transfer rule. In both problems, the minimizing truthful total transfers are unique up to
agent-specific constants summing to zero and events of probability zero.
Read the full abstract ▾
Computation and Language
PHRBench: A Behavioral Evaluation of Post-Hallucination Reasoning in LLMs
Linghao Meng, Feng He, Xuan Yang et al. (+5 more)
Hallucinated information can propagate through multi-stage LLM systems and become part of the context for subsequent
reasoning. Existing studies of post-
hallucination reasoning (PHR) mainly characterize changes in final outcomes and aggregate
reasoning dynamics, leaving how models resolve hallucinated premises at the response level insufficiently understood. In this work, we introduce PHRBench, a controlled
benchmark for behaviorally structured PHR across four domains and 18 large
language models. PHRBench characterizes each
reasoning trajectory independently of final-answer correctness through
Hallucination Compliance,
Hallucination Avoidance, and Heuristic Correction, and defines an insightful trajectory as successful correction that ultimately reaches the correct answer. Across 4820 controlled instances, we find that successful recovery remains relatively rare and is associated with more frequent belief updates along the
reasoning trajectory. We further find that properties of the hallucinated
prompt contain substantial predictive signal for successful recovery, with a lightweight predictor achieving an AUROC of 0.847. These findings provide a behavioral view of post-
hallucination reasoning, characterizing how LLMs resolve erroneous context and when successful recovery is likely to occur.
Read the full abstract ▾
All authors Linghao Meng, Feng He, Xuan Yang, Junyuan Mao, Pinze Ren, Deqing Mu, Hesen Yang, Qiankun Li
Strongly Correlated Electrons
Resolving Spectral Gaps of the Frustrated Spin-1/2 $J_1$-$J_2$ Chain via Dual Spin-Sector Variational Quantum Eigensolver
Sk Saniur Rahaman
Finding excited states and resolving spectral gaps in frustrated
quantum spin systems using Variational
Quantum Eigensolvers is complicated by state-ordering inversions. In the frustrated
spin-1/2 $J_1$-$J_2$ Heisenberg chain, standard single-sector variational
quantum deflation collapses onto intermediate triplet excitation, preventing the isolation of excited singlets and the singlet-singlet gap. To resolve these gaps, a Dual
Spin-Sector Variational
Quantum Eigensolver framework is introduced for near-term
quantum architectures. By partitioning the search across magnetization sectors $S^z=0$ and $S^z=1$, the lowest triplet is resolved as an unconstrained ground state in $S^z=1$ with zero deflation penalty. Leveraging exact global
spin-rotation
symmetry, this state is mapped analytically into $S^z=0$ without ancilla overhead, unlocking dual deflation via shallow compute-uncompute
circuits. Tracking both singlet-triplet and singlet-singlet gaps across coupling ratios $J_2/J_1$ from $0$ to $1$ for chains up to $16$
qubits faithfully captures all critical phenomena: the continuous Berezinskii-Kosterlitz-Thouless transition, exact Majumdar-Ghosh dimer degeneracy, level-ordering inversion in the gapless
liquid phase, and incommensurate spiral excitations.
Circuit depth scaling up to twenty layers provides an overparameterization mechanism overcoming classical local minima in strongly frustrated regimes, reducing relative energy errors below $1\%$ and elevating state fidelities above $97\%$ on $16$
qubits. Projective measurement sampling demonstrates resilience against shot noise, providing a practical blueprint for exploring frustrated
quantum magnetism on near-term
quantum devices.
Read the full abstract ▾
Robotics
RFPO: Rectified Flow Policy Optimization for Embodied Control
Ting Huang, Lisiyu Pan, Haoyu Wang et al. (+6 more)
Flow-based policies provide an expressive framework for continuous
robot control, but their iterative ODE integration incurs substantial inference cost. Naively reducing the integration budget can severely degrade control, since policies optimized under full-step execution are not explicitly constrained to remain reliable under coarse numerical integration. We refer to this mismatch as the few-step discretization gap. To address this problem, we introduce RFPO, a flow-policy optimization framework for reliable few-step execution. Reward-aware online Reflow rectifies student-induced transport paths during on-policy learning, making the resulting policy more robust to coarse integration. A frozen Gaussian PPO controller supplies complementary action-space supervision at full and intermediate integration budgets, while the deployed policy remains a single flow student executed with one Euler step. Across Unitree Go2, Boston Dynamics Spot, Unitree H1, and Unitree G1, RFPO consistently preserves full-step control performance under one-step execution, with one-step returns remaining within 2.4% of their corresponding 64-step values across both zero and random initialization. On Unitree Go2, one-step execution retains 98.5% of the 64-step reward while reducing onboard mean inference latency from 4.39 ms to 0.08 ms, yielding a 54.9x speedup. Real-
robot experiments further validate stable one-step locomotion. Code: https://github.com/AIGeeksGroup/RFPO. Website: https://aigeeksgroup.github.io/RFPO.
Read the full abstract ▾
All authors Ting Huang, Lisiyu Pan, Haoyu Wang, Zeyu Zhang, Siyuan Qian, Yanjun Li, Yandong Guo, Boxin Shi, Hao Tang
Number Theory
$p$-class groups in the cyclotomic $\mathbb{Z}_p$-extension of imaginary quadratic fields
Somnath Jha, Debanjana Kundu, H. Laxmi et al. (+1 more)
We study the structure of the $p$-class group of the first layer of the cyclotomic $\mathbb{Z}_p$-extension of an imaginary quadratic field where $p$ is non-split. We obtain restrictions on the growth of the cyclic factors from the base layer to the first layer; in particular, when the $p$-rank remains unchanged, each cyclic factor grows by exactly one power of $p$. We show that in a large number of cases the Iwasawa $\lambda$-invariant is completely determined by the $p$-rank of the first layer itself. Finally we use a Cohen-Lenstra-Martinet type philosophy to study the probability distribution of Iwasawa lambda invariants of the first layer when the base layer has cyclic $p$-class group.
Read the full abstract ▾
All authors Somnath Jha, Debanjana Kundu, H. Laxmi, Lawrence C. Washington
Algebraic Geometry
Combs and complex structures on the six-sphere
Jeff Viaclovsky
We construct complex structures on the six-sphere $S^6$ with arbitrarily large local Kuranishi dimension. As a corollary, we find the same phenomenon holds for the smooth
manifold underlying complex projective 3-space $\mathbb{P}^3$.
Fluid Dynamics
Phase-controlled lateral forcing of droplets on a vibrating bath
Lawrence A. Wilen, Zahra Shamsi, Johnathan Hoggarth et al. (+1 more)
We introduce a method for controlling the lateral motion of droplets bouncing on a vibrated
liquid bath by superimposing a small periodic rocking motion onto the conventional vertical forcing. We show experimentally that the lateral velocity varies linearly with the rocking amplitude $\alpha$ and sinusoidally with its phase $\phi$ relative to the vertical forcing, while the response depends strongly on the droplet's vertical bouncing mode. A time-averaged model of a harmonic bouncer yields a time-averaged lateral force $\overline{F}_x=mg\alpha\sin(\Phi_b+\phi)$, where $\Phi_b$ is the droplet impact phase. Periodic rocking thus provides a simple, programmable means of exerting lateral forces on bouncing droplets, enabling controlled manipulation of passive and active droplets in pilot-wave hydrodynamics.
Read the full abstract ▾
All authors Lawrence A. Wilen, Zahra Shamsi, Johnathan Hoggarth, Bauyrzhan K. Primkulov
Fluid Dynamics
SAGE-PINN: A Singularity-free Axisymmetric Geometry-Encoded Physics-Informed Neural Network for Axisymmetric Multiphysics Flow in Cylindrical Coordinates
Raj Maurya, Joyprakash Akhuli and Doyel Pandey
Physics-informed neural networks (PINNs) provide a mesh-free framework for multiphysics modeling, yet fluid flow in axisymmetric cylindrical coordinates remains largely unexplored because the governing operators are singular at the axis and momentum and energy are strongly coupled through advection and buoyancy. We introduce SAGE-PINN, a Singularity-free Axisymmetric Geometry-Encoded Physics-Informed Neural Network for this class of problems. SAGE-PINN removes the coordinate singularity by construction: the parity-preserving variable $s=r^2$ and a modified stream-function representation eliminate explicit $1/r$ terms from the trained residuals and satisfy mass conservation identically. Boundary-fitted analytical ansätze enforce the boundary conditions exactly, while staged momentum-thermal-joint training, residual normalization, Péclet- and Richardson-number continuation, and a physics-scaled drift guard stabilize the coupled optimization. The framework is trained without labelled data, using only physics and boundary constraints, and its predictions are validated against finite-element reference solutions for steady nanofluid flow and heat transfer through a stenosed artery over a wide range of flow and thermal regimes, without case-specific tuning. Mass conservation remains at single-precision round-off, and over the central operating range the axial velocity and temperature are reproduced within $1.6\%$ and $9.1\%$ relative $L_2$ error, respectively. The main deterioration occurs only in the limiting regimes of strongly buoyancy-influenced slow flow and highly advection-dominated heat transfer.
Embedding cylindrical geometry, conservation structure, and boundary physics directly into the network provides a robust route to PINN simulation of axisymmetric multiphysics flows.
Read the full abstract ▾
Multimedia
MemoCare: An Interactive Multimodal Mobile System for Automated Cognitive Screening
Duy-Cat Can, Mau Minh Phuc Le, Tuan-Khoa Hoang et al. (+10 more)
MemoCare is an interactive mobile system for automated
multimodal cognitive screening. A React Native application combines spoken responses, temporal and spatial orientation, touchscreen actions, and visuoconstruction in complete English and Vietnamese workflows.
Speech is transcribed by Google
Speech-to-Text and scored locally with deterministic task-specific natural
language processing rules; GPS coordinates are resolved by the MemoCare spatial module before answer matching; touch tasks are scored from interaction events; and the drawing task uses a three-model convolutional neural network consensus with separate visual interpretation. Software tests pass 151/151 predefined cases across
speech/
language, spatial-answer, and touch-interaction scoring, while spatial regression passes 48/48 four-country coordinate-resolution cases. For the drawing module, validation-selected ShuffleNetV2 x1.5 achieved 91.33% mean balanced accuracy and 78.87% exact three-criterion accuracy on a locked 71-image test set. Four clinician co-authors additionally inspected the end-to-end workflow, yielding a pooled median rating of 4/5 across eight criteria, with item-level medians ranging from 3 to 4.5. At MMM, attendees can directly try a shortened
multimodal screening workflow and inspect automatic item-level and total scoring.
Read the full abstract ▾
All authors Duy-Cat Can, Mau Minh Phuc Le, Tuan-Khoa Hoang, Hai-Dang Nguyen, Trung-Hieu Do, Dang Minh Ly, Minh-Duc Nguyen, Nghia TT Hoang, Linh-Trung Nguyen, Huy-Hieu Pham, Huong Ha, Binh T. Nguyen, Oliver Y. Chén
Machine Learning
A Good Self-Teacher Meets the Student Where They Are: Joint On-Policy Learning and Teaching
Randy Ardywibowo, Arnav Dalal and Jiantao Jiao
Reinforcement Learning (RL) from outcome rewards suffers from sparse supervision, particularly on difficult, long-horizon tasks where successful trajectories are rare and costly to generate. On-Policy
Distillation (OPD) offers an attractive alternative by providing dense
token-level supervision from a stronger teacher along the student's own generations. Self-
distillation methods further remove the need for a separate teacher model by conditioning the same policy on privileged information to serve as its own teacher. However, privileged conditioning alone does not guarantee that the resulting
distillation update improves the student. Indeed, privileged information can lead the teacher to solve tasks through shortcuts unavailable to the student, producing supervision poorly matched to the student's current behavior. Consequently, even a higher-performing teacher can provide guidance that degrades student performance. To address this, we analyze how the choice of privileged teacher affects the student's update. We derive a necessary and sufficient condition for the teacher's local
distillation update to be a positive multiple of the student's reward gradient. Our analysis suggests that the teacher should not only perform well on the task, but also provide guidance suited to the student's current capabilities. This characterization motivates a practical teacher-training surrogate that combines outcome rewards with
token-level Kullback-Leibler (KL) regularization toward the student. Based on this result, we propose Joint On-Policy Learning and Teaching (JOLT), which jointly trains a single policy in two roles: a privileged teacher using a KL-regularized objective, and an unprivileged student using dense on-policy
distillation. Across mathematical
reasoning, coding, tool use, and terminal use, JOLT improves training efficiency and performance, with further gains from student rewards.
Read the full abstract ▾
Data Structures and Algorithms
Subset Sum via Partial Match
Lixi Ye and Baitian Li
We show that the subset sum problem can be solved in time $O(2^{0.499999n})$, breaking the $2^{n/2}$ meet-in-the-middle barrier. Our approach builds on the representation technique framework of Randolph and Węgrzycki (STOC 2026). We choose a representation for which testing the compatibility of pairs of partial solution vectors is exactly the partial match problem. To beat exponent $1/2$ for subset sum, it then suffices to give a nontrivial partial match
algorithm in a certain parameter regime. We achieve this by designing a depth-2 linear
circuit for the partial match
matrix, which yields an efficient
algorithm via the framework of Alman and Li (FOCS 2025).
Read the full abstract ▾
Complex Variables
Gromov hyperbolicity, finite type, and subellipticity on $\mathbb{C}$-convex domains
Amar Deep Sarkar
In this article, we study three properties of a bounded $\mathbb{C}$-convex domain $\Omega \subset \mathbb{C}^n$, $n \geq 2$, with smooth boundary, namely Gromov hyperbolicity of its Kobayashi distance, finiteness of the D'Angelo type of $\partial \Omega$, and the existence of a subelliptic estimate for the $\bar{\partial}$-Neumann problem on $(0,1)$-forms. We prove that these three properties are equivalent. The subelliptic estimates are obtained without Catlin's construction of plurisubharmonic weights. Our weights are explicit bounded transforms of the logarithm of the Bergman kernel. A single quantity controls all three properties, namely the radius of the largest disc through a point of $\Omega$ in a given complex direction. Gromov hyperbolicity forces a power bound for these radii. Finite type gives the power $1/M$, where $M$ is the maximal type, and this power cannot be improved. A subelliptic estimate gives a power bound through the canonical solution operator applied to normalized Bergman kernels. The analytic estimates need no boundary regularity. On a bounded pseudoconvex domain, a bounded weight whose induced Hessian on $(0,q)$-forms is bounded below by a negative power of the boundary distance gives a Sobolev estimate for the extension by zero of $(0,q)$-forms, and also
eigenvalue bounds for the $\bar{\partial}$-Neumann operator. On smooth $\mathbb{C}$-convex domains of finite type, we show that the directional expansion exponent, the normal expansion exponent, and the supremum of the subelliptic gains are all equal to $1/M$. Next, we prove that on a Lipschitz $\mathbb{C}$-convex domain whose Kobayashi distance is Gromov
hyperbolic, the Euclidean boundary and the Gromov boundary are bi-Hölder equivalent. Finally, we give a Hardy-type criterion for Gromov hyperbolicity of collar metrics.
Read the full abstract ▾
Artificial Intelligence
RunningTab: Direct Workspace Interaction with Environment-Side Tabs
Jinheon Baek, Soyeong Jeong, Yumin Choi et al. (+2 more)
Much knowledge work produces new deliverables from files a workspace already holds, and LLM
agents are beginning to take such work over. Through direct corpus interaction, an
agent can search and read any of those files from a terminal with no indexing, and producing a deliverable from many of them in this way is what we call direct workspace interaction (DWI). Reaching the files, however, is only half the task: nothing keeps track of what the task asks for, what has been read, and what was listed but never opened, all of which slip through the context window without leaving a trace, so an
agent may extract a figure and still deliver a report without it. To address this, we present RunningTab, a framework that equips direct workspace interaction with an environment-side tab: a per-task record of what the task still owes, kept by the environment alongside the
agent. Specifically, the
agent adds its requirements, while the environment records every file read as an excerpt with its provenance and every listed but unopened file as a candidate; the
agent can then see each requirement beside its best-matching excerpts and top unopened candidates, resolve it against matching content or set it aside with a reason, and, should it try to finish with requirements still open, receive them in a finish check. We validate RunningTab on three
benchmarks with three LLMs, where it consistently outperforms plain DWI and baselines that keep the record in the model, while its tab usually holds the values a deliverable needs once seen.
Read the full abstract ▾
All authors Jinheon Baek, Soyeong Jeong, Yumin Choi, Dongsu Han, Sung Ju Hwang
Data Structures and Algorithms
Color Coding for the Sherrington-Kirkpatrick Model
Alina Harbuzova, Saba Lepsveridze, Mahbod Majid et al. (+1 more)
We give a
polynomial-time
algorithm for approximating the partition function of mean-field mixed $p$-
spin models to arbitrarily high accuracy throughout the second-moment regime. This improves the quasipolynomial-time
algorithm of Bencs, Huang, Lee, Liu, and Regts (arXiv:2507.15616) and answers their open question. In particular, our result covers the entire replica-symmetric regime of the Sherrington–Kirkpatrick model. The
algorithm is deterministic, runs in time
polynomial in $n$ and $1/\varepsilon$, and succeeds for every typical realization of the disorder. Our main technical contribution is a new algorithmic application of color coding to the classical high-temperature expansion introduced by Aizenman, Lebowitz, and Ruelle in their study of fluctuations of the Sherrington–Kirkpatrick partition function. We combine this approach with the zero-freeness established by Bencs et al. (arXiv:2507.15616) to convert additive approximations into multiplicative ones.
Read the full abstract ▾
All authors Alina Harbuzova, Saba Lepsveridze, Mahbod Majid, Ankur Moitra
Superconductivity
Correlating DC SQUID Performance with the Location of Trapped Magnetic Flux Using Scanning SQUID Microscopy
Bochao Xu, Ian W. Haygood, Kyle Jackman et al. (+3 more)
Holes in the ground planes of superconducting
circuits, commonly referred to as moats, are known to be an effective means of mitigating the deleterious effects of residual
magnetic flux in superconducting
circuits. Previous studies have utilized scanning SQUID microscopy (SSM) to image the effectiveness of various moat geometries and the maximum
magnetic fields where they no longer prevent
vortices in the ground plane but have not directly correlated the location of flux with
circuit performance measurements. In this study we employ SSM to image the effectiveness of various moat configurations in trapping unwanted flux near a DC SQUID, and when the moats are not effective, the location of
vortices in the ground plane. During the same cooldown, and without disturbing the
circuit, we measure the
circuit performance to directly correlate this with the
vortex location. We show that there are preferential pinning sites for fluxons regardless of moat configuration, demonstrating the utility of the SSM in provide feedback for
circuit layout, modeling, and failure analysis.
Read the full abstract ▾
All authors Bochao Xu, Ian W. Haygood, Kyle Jackman, Coenrad J. Fourie, Pete F. Hopkins, Michael L. Schneider
Social and Information Networks
CrossWeave: Bridging Perspectives Across Online Communities with a Dual-Pane Design
Fei Fang, Reva Hirave, William Jurayj et al. (+19 more)
Social media systems typically display conversations among already familiar contributors, which can be predictable and one-sided. In civic discourse, this design narrows discussion, reinforces divides, and distorts the perception of public opinion. To encourage cross-community engagement, we present CrossWeave, an AI-powered bridging system that augments the standard social media feed. As the user reads a post, CrossWeave surfaces diverse relevant posts from other threads in a side pane and highlights the connections. Users are invited to venture out of their echo chamber, explore a broader range of views and arguments, and “click across” to engage with their authors. When they do, CrossWeave facilitates constructive posting, not only by showcasing relevant past content but also by simulating possible reactions as the user drafts a post.
Read the full abstract ▾
All authors Fei Fang, Reva Hirave, William Jurayj, Yuqi Li, Brian Lu, Tarik Metin, Tsugunobu Miyake, Kateryna Morhun, Yash Permalla, Kenan Rustamov, Allen Shen, Haojun Shi, Prabhav Singh, Xiheng Tom Wang, Kevin Xu, Qingcheng Zeng, Jiayi Zhang, Daniel Khashabi, Andrew Perrin, Tiziano Piccardi, Ziang Xiao, Jason Eisner
Machine Learning
Seq-Flow: Efficient Probabilistic Forecasting with Self-Rollout Error Control
Yinan Huang, Shitij Govil, Bo Dai et al. (+1 more)
Many scientific
forecasting tasks require updating a distribution over future trajectories as new observations arrive. Conventional
diffusion and flow models generate each forecast from Gaussian noise, often at the cost of many sampling steps. Warm-start methods reuse earlier predictions to reduce this cost, but their models are not trained to perform the forecast update itself, which can compromise quality under few-step sampling. In this work, we introduce Seq-Flow, a conditional flow model whose ODE transports samples from the previous forecast distribution to the updated one. Because successive forecasts often differ only modestly, this transport starts from an informative distribution and can produce accurate updates with few flow evaluations. Recursive reuse also creates a challenge: errors in one forecast become errors in the initial states of subsequent flows. We address this with self-rollout training, in which a moving average copy of the model generates forecasts that initialize later training updates. Unlike self-forcing methods, which reuse generated outputs as conditioning context, Seq-Flow reuses them as the source of the next flow. Experiments On particle-accelerator beam spill
forecasting show Seq-Flow reduces CRPS by 65% under a few-NFE sampling budget, while remaining competitive with strong baselines on fluid-dynamics
forecasting tasks. Although trained on self-rollouts of at most four updates, Seq-Flow remains accurate over more than 400 consecutive updates. Our code is available at https://github.com/Graph-COM/Seq-Flow.
Read the full abstract ▾
All authors Yinan Huang, Shitij Govil, Bo Dai, Pan Li
Algebraic Geometry
Abundance for threefolds in positive characteristic
Fabio Bernasconi, Jakub Witaszek and Zheng Xu
We prove the abundance
conjecture for threefolds in characteristic $>42$.
Computer Vision
ECHO: Embodied Camera Observations of Human Object Carrying
Xuefei Sun, Lorin Achey, Kali Hamilton et al. (+4 more)
Embodied and assistive
agents must do more than recognize objects: they must reason about where an object belongs given the layout of an environment and the habits of the people who live in it. Progress on this problem has been limited, in part because no dedicated
benchmark or
dataset exists to define and evaluate it. Existing RGB-D scan
datasets reconstruct static rooms without human activity, while human-object-interaction
datasets capture motion without a navigable, fully reconstructed scene or a ground-truth notion of an object's natural destination. We introduce contextual object placement as a
benchmark task: predicting an object's destination during an observed object-carrying episode. To support this task, we present
Embodied Camera observations of Human Object carrying (ECHO), a large-scale synthetic
dataset that pairs dense RGB-D scans of indoor scenes with recordings of an
embodied human carrying everyday objects to context-appropriate destinations. ECHO is the first publicly available
dataset to combine reconstructed scenes, human activity, natural
language, and contextual-placement annotations. It comprises 3,805 human-annotated episodes across 159 floors of 115 HM3D scenes, involving 198 distinct objects. Each floor includes a complete RGB-D scan with human-annotated room labels and a surface list. Each episode provides synchronized RGB-D encounter clips; 6-DoF camera, human, and object trajectories; start and destination surfaces; an action caption; and a human-written context: a single sentence describing the inhabitant's routine that implies the destination without naming it. We evaluate contextual object placement using input-masked probes and an end-to-end baseline. Results show that no single input modality is sufficient, highlighting the need to jointly reason over scene structure, human activity, and contextual knowledge.
Read the full abstract ▾
All authors Xuefei Sun, Lorin Achey, Kali Hamilton, Alberto Speranzon, Gregory Grebe, Yonatan Bisk, Christoffer Heckman
Machine Learning
Q-Learning with Scalar Adjoint Matching
Yonghoon Dong, Minsung Yoon, Jaehyuk Kim et al. (+3 more)
Flow policies capture rich and diverse action distributions, and fine-tuning them with off-policy RL to improve beyond the demonstrations has drawn growing interest. However, fine-tuning a flow policy against a learned value function is not trivial, because the policy generates its action over many flow steps. Adjoint matching offers a principled way to update the flow model itself by propagating value information from the final action back to each flow step, but it requires a vector–Jacobian product through the policy at every step, a cost that grows with the number of flow steps and the policy size. We observe that the batch-averaged velocity Jacobian of pretrained flow policies concentrates on its diagonal. Motivated by this finding, we derive a closed-form scalar adjoint that scales the value gradient at the final action by the flow time, eliminating the per-step vector–Jacobian products. We further find that controlling the critic's value at policy-generated actions is particularly important under the scalar adjoint. Based on these findings, we propose Q-learning with Scalar Adjoint Matching (SQAM), which combines the scalar adjoint with a value penalty at those actions. SQAM's gains concentrate on the four hardest OGBench domains, where its success rate exceeds that of the strongest baseline in each domain by 18 to 35 percentage points. To test whether SQAM extends to large pretrained policies, we also fine-tune a
vision-
language-action policy on a real bimanual
robot. SQAM improves over supervised fine-tuning on all three tasks.
Read the full abstract ▾
All authors Yonghoon Dong, Minsung Yoon, Jaehyuk Kim, Jungwoo Park, Changyeon Kim, Jinwoo Shin
Computer Vision
Detecting Adversarial Images through Response Profiles of Vision-Language Models
Arash Vashagh and Roozbeh Razavi-Far
Adversarial perturbations can alter the predictions of frozen
vision-
language models (VLMs) while leaving their confidence and image–text similarity patterns seemingly plausible. We investigate whether we can identify
adversarial inputs based on the broader way an image interacts with a collection of general semantic
prompts. Our detector summarizes these responses using category-level statistics, relationships among
prompts, deviations from clean reference distributions, and stability under weak image transformations, producing a compact response profile that is classified by a lightweight model while the VLM remains fixed. We evaluate the approach on multiple public image
datasets, several CLIP-style visual backbones, and a range of gradient-based, optimization-based, automated, and spatial attacks. The detector achieves strong discrimination in attack-specific settings and retains substantial performance when evaluated on attacks not seen during training. Under a controlled detector-specific protocol, the response-profile representation outperforms the evaluated
embedding-geometry baselines. Additional analyses show that the feature groups provide complementary information and that the method remains effective under variations in the
prompt configuration. We also examine inference cost and performance against detector-aware adaptive attacks. Overall, the results indicate that response patterns across semantic
prompts provide a useful complementary signal for
adversarial image detection in frozen VLMs.
Read the full abstract ▾