Papers

With the rapid development of quantum computers, proofs of quantumness have recently become an interesting research direction. However, in current schemes for proofs of quantumness, quantum provers face the risk of being maliciously exploited by classical verifiers. Through malicious strategies in interaction with quantum provers, classical verifiers could solve some instances of hard problems that arise from the specific scheme in use. This is due to the lack of formalization that prevents malicious verifiers from extracting useful information in proofs of quantumness. To address this issue, we formalize zero-knowledge proofs of quantumness. Intuitively, the zero-knowledge property necessitates that the information gained by the classical verifier from interactions with the quantum prover should not surpass what can be simulated using a simulated classical prover interacting with the same verifier. As a result, the new zero-knowledge notion can prevent a malicious verifier from exploiting quantum advantage. We find that the classical zero-knowledge proof is sufficient to compile some existing proofs of quantumness schemes into zero-knowledge proofs of quantumness schemes. It appears to be more general to require zero-knowledge proof on the verifier side instead of the prover side. This helps to regulate the verifier's behavior from malicious to be honest-but-curious. As a result, both parties will play not only one role in the proofs of quantumness but also the dual role in the classical zero-knowledge proof. Specifically, Shor's factoring-based scheme and the learning with errors-based scheme in [Brakerski et al., FOCS, 2018] can be transformed into zero-knowledge proofs of quantumness by requiring an extractable non-interactive zero-knowledge argument on the verifier side. Zero-knowledge proofs of quantumness can thus be viewed as an enhanced security notion for proofs of quantumness.

Two groups of autonomous, anonymous, and oblivious mobile robots are deployed in the two-dimensional Euclidean plane, each assigned a distinct task. We study a setting where the two groups must simultaneously solve two conflicting pattern formation problems: the gathering problem, where robots gather at a point not known to them a priori, and the circle formation problem, where robots occupy distinct positions on the boundary of a circle. Although each robot knows its own task, it cannot identify other members of its group. A prior solution~\cite{Conflict-1} addressed this problem for asynchronous robots having direction-only axis agreement and global weak multiplicity detection capability available to all robots in both groups. In contrast, in this work, we consider fully disoriented robots without any axis agreement or common chirality. We study the feasibility of a solution to this problem for disoriented robots. We propose a distributed algorithm that solves the problem for semi-synchronous disoriented robots with non-rigid movements. Our proposed algorithm assumes global weak multiplicity detection only for the gathering group, while for the circle formation group, it requires local weak multiplicity detection.

Cinematic audio source separation (CASS) decomposes a soundtrack into dialogue, music, and sound-effects (SFX) stems. Existing CASS methods, however, suffer from two critical limitations: they rely on heavily parameterized network architectures and GPU-class hardware, limiting their use in real-time and resource-constrained scenarios, and they are overwhelmingly designed for monaural signals, leaving the stereo scenario largely unexplored. We present LiteCASS, to our knowledge the first lightweight end-to-end network for real-time stereo CASS. LiteCASS combines deterministic STFT subband rearrangement with two jointly trained compact U-Nets: the first extracts dialogue, and the second separates music and SFX from the predicted non-speech component. A multi-task waveform-domain L1 loss supervises all stems. On a spatialized stereo extension of DnR v3, LiteCASS-K8 uses only 1.06M parameters and 0.72G MACs per second, while achieving the highest averaged SI-SDR among the compared CASS baselines.

Heavy-quark production in deep-inelastic scattering probes the gluon distribution, and the precision of current and future data requires predictions beyond next-to-leading order. We calculate the even Mellin moments $N = 2$ to $22$ of the next-to-leading-order heavy-quark coefficient functions for $F_2$ and $F_L$, retaining the full heavy-quark mass dependence. This includes the gluon channel, for which no analytic result is available. They provide a test of the parametrisations used in phenomenology: at a representative kinematic point these are accurate to better than one per cent in the gluon channel, whereas in the quark channels we find deviations of up to $20\%$ over a wide range of Mellin moments, signalling significant differences in the corresponding $x$-space distributions. As a new result at next-to-next-to-leading order, the lowest moment $N = 2$ has been determined exactly.

Social influence can preserve errors that individuals would correct on their own. We study this possibility by representing opinions as directions in a semantic space. The formulation connects established models of personal preference, social agreement, and external influence. A common factual signal pulls every opinion toward a reference direction, individuals remain attached to their initial preferences, and connected individuals imitate one another. When the factual signal exceeds the preference strength, each isolated individual corrects, apart from an exactly opposite initial opinion. We prove that every finite undirected network then has a unique state of minimum total energy in which all opinions favor the factual direction. This state is the best overall compromise between factual alignment, personal preferences, and agreement with neighbors. Yet on a ring with planar opinions, sufficiently strong imitation can trap the same individuals in a different stable state that retains misinformation. We prove convergence to that state from the specified initial opinions. We also derive a network spectral condition that excludes competing stable states, and show how an additional semantic direction can destabilize the constructed planar pattern. Numerical maps reveal persistent error, history dependence, and coexisting outcomes, while comparisons across networks show differences that a single spectral scale does not remove. These results distinguish individual correctability from collective correction and identify how attractive social influence can obstruct the latter.

Accurate cue-ball identification is essential for metric analysis of broadcast snooker. Existing systems evaluate each candidate independently against a fixed white prototype and reject candidates above an appearance threshold. Under broadcast conditions, illumination changes can make colored balls appear white, while intrusions from players and equipment can obscure the cue ball or introduce competing candidates. We formulate cue-ball identification as a rule-constrained assignment problem that jointly assigns detected candidates to the bounded snooker inventory: one cue ball, up to 15 reds, and six colors with known spots. The cue ball is selected by the incremental cost of assigning each candidate to the white slot, and the conventional appearance test follows as the one-slot case. On 419 hand-annotated shots, our method improves identity accuracy from 88.1% to 95.5%, and from 80.5% to 95.2% on held-out venues. Within CueLift, our metric state-recovery system, the assignment expands coverage from 36.2% to 55.4% over 6,241 scorable shots. Assigning an estimate to every shot in a separate evaluation on 2,529 shots preserves this advantage.

Memory systems are becoming a core component of LLM agents, but constructing and maintaining memory remains expensive because it relies on repeated calls to large proprietary language models. This cost creates a major barrier to deploying memory-enhanced agents at scale. In this paper, we present Pseudo Self-Distillation (PSD), a framework that enables small language models (SLMs) to construct hierarchical memory representations by distilling behavior from a strong black-box oracle through a multi-stage training pipeline. Standard distillation methods require access to teacher logits or hidden states, which closed models do not expose. Unlike conventional self-distillation settings, where supervision is derived from a model's own predictions, sampled rollouts, or aggregated outputs, PSD enables a single-model distillation setup while channeling external oracle knowledge through the prompt. PSD uses a single small model in two roles: a teacher that sees a privileged prompt containing the oracle's answer as reference context, and a student that sees only the task prompt. The student learns to reproduce the teacher's output distribution, absorbing oracle-guided behavior into its own weights without accessing the oracle's internals. On LoCoMo, PSD-trained Qwen3-0.6B, 1.7B, and 4B match or exceed GPT-4.1-mini on downstream retrieval at a fraction of the deployment cost, with off-policy PSD achieving the strongest results across most conditions. We further show that this memory-construction capability transfers out-of-distribution to LongMemEval, despite the students being trained exclusively on LoCoMo with no exposure to LongMemEval data.

We study the nonlinear stability of a composite wave consisting of a traveling wave and a rarefaction wave for a parabolic-hyperbolic system arising from chemotaxis. We prove that if the initial value is a small $H^1$-type perturbation of composite wave, then the system admits a global solution that converges toward the composite wave with an absolutely continuous shift. The proof combines the weighted relative-entropy mechanism for viscous shocks with the energy structure of rarefaction waves. A key ingredient is the inclusion of a rarefaction modulation factor in the weighted relative entropy; its derivatives combine with the terms generated by the rarefaction profile to produce a rarefaction dissipation. Moreover, the spatial separation of the two waves yields time-integrable interaction errors caused by the non-exact superposition. Since diffusion acts only on the density component, the full $H^1$ estimate is closed by exploiting the coupling structure of the system to recover the missing dissipation of the hyperbolic component.

Metal-mediated growth provides an effective route to high-quality hexagonal boron nitride (hBN), but hBN formation on the metal does not necessarily ensure hBN formation on an adjoining dielectric substrate. Here, we investigate this distinction using a vertically stacked Ni–Cr/boron/sapphire configuration operated at 1550~$^\circ$C under atmospheric pressure, with the relative boron supply varied at fixed Ni:Cr = 8:2. hBN is detected by Raman spectroscopy on the gas-exposed alloy surface throughout the investigated Alloy:B range, whereas sapphire-side hBN appears only at the higher relative boron supplies of Alloy:B = 24:1 and 18:1, as confirmed by both Raman spectroscopy and XRD. The hBN (002) reflection further narrows as the boron supply increases from 24:1 to 18:1, corresponding to an increase in the apparent c-axis coherent length from approximately 86 to 143~nm. Additional crystalline reflections also emerge only under these higher-boron conditions. These results identify boron supply as a key parameter governing the transition from alloy-side hBN formation to concurrent sapphire-side hBN growth. Once formed, sapphire-side hBN occurs within a substantially reconstructed and chemically modified Al/O-rich sapphire environment rather than as a uniform film on an intact substrate. To our knowledge, this study provides the first demonstration of successful sapphire-side hBN formation by Ni–Cr-mediated growth.

Let $N\geq 3$, $2<q<p<2^*$, and let $V\in L^\infty(\mathbb{R}^N)$ be real-valued and $\mathbb{Z}^N$-periodic. We assume that $0$ belongs to a finite spectral gap of the periodic Schrödinger operator $L=-\Delta+V$. We prove that there exists $\lambda_\infty>0$ such that, for every $\lambda\geq\lambda_\infty$, the competing-power equation $$ -\Delta u+V(x)u=|u|^{p-2}u-\lambda |u|^{q-2}u \quad\text{in }\mathbb{R}^N $$ has a nontrivial solution $u_\lambda\in H^1(\mathbb{R}^N)$. Moreover, $$ \|u_\lambda\|_{H^1(\mathbb{R}^N)}+\|u_\lambda\|_{L^\infty(\mathbb{R}^N)} \leq C\lambda^{-1/(q-2)}. $$ We also show that, after the natural rescaling and lattice translations, a sequence of such solutions converges weakly to a nontrivial solution of the pure defocusing equation $Lv=-|v|^{q-2}v$.

Bimanual manipulation requires policies that coordinate two arms while adapting their functional roles to scene geometry, object configuration, and task context. Learning such scene-conditioned role adaptation remains challenging, as demonstrations may contain uneven role distributions that limit generalization to underrepresented arm–role configurations. In addition, many bimanual policies predict actions in fixed left- and right-arm action spaces. While this provides a natural parameterization for robot control, it does not explicitly specify how behaviors should transform when functional roles are exchanged across arms. Across different scene initializations, the two arms may follow a similar coordination pattern, but the role-specific behavior assigned to each arm should change with the scene. Therefore, we propose BiRoAD, a Bimanual Role-Adaptive Decomposition framework for learning shared and role-adaptive representations in bimanual policies. Given bimanual trajectory or action-token features, BiRoAD decomposes these features into swap–symmetric and swap–antisymmetric components: the former captures coordination structure invariant to arm exchange, and the latter captures role-specific distinctions that vary consistently with functional role assignment. The two components are then recomposed as residual updates to the original paired arm representations, allowing BiRoAD to serve as a modular feature transformation without changing the policy inputs, imitation-learning objective, or requiring manually defined role labels. Across multiple bimanual manipulation tasks with balanced and imbalanced role distributions, BiRoAD improves robustness across role configurations over corresponding base policies, with notable gains on underrepresented role configurations.

Engineering change order (ECO) is an important step in repairing timing and electrical violations during the late stages of chip design. Existing Agentic EDA methods primarily focus on tool invocation, with less attention to model decision quality and targeted training. A central challenge in ECO is multi-round decision-making: the model must use the results of each round to determine the next repair action. We propose WaveletECO, which integrates a closed-loop execution platform with large language models to enable agents to execute ECO decisions effectively. We also train a local 9B model through supervised fine-tuning and CPO-SimPO using execution demonstrations and decision-preference data, enabling ECO decision-making with a locally deployed model. Across 594 evaluation runs on 22 designs, WaveletECO-Policy (BF16) and (INT8) score 79.63 and 79.65, respectively, compared with GPT-6 Astra's 77.44. The estimated inference cost of INT8 is about 1/147 of GPT-6 Astra's. These results show that specialized model training supports effective, low-cost multi-round ECO repair, with repair quality retained under INT8 quantization.

Let $(K^h,v^h)$ be a Henselization of a valued field $(K,v)$, $\mu$ any extension of $v$ to $K[x]$, and $\mu^h$ the canonical extension of $\mu$ and $v^h$ to $K^h[x].$ In this paper, we study two new invariants associated with a Mac Lane-Vaquié chain (MLV) of $\mu$, namely, the defect of each augmentation step and the regularity of an MLV chain of $\mu$. This, in turn, extends the notion of regularity of a complete sequence of abstract key polynomials for $\mu$ to the regularity of $\mu$ itself. We also prove that the regularity of $\mu$ is a sufficient condition for the equality of the depths of $\mu$ and $\mu^h$. Furthermore, it is a necessary and sufficient condition for the equality of the defect and relative gap of each augmentation step of an MLV chain of $\mu.$ Finally, we show that if $\mu$ has finite degree, then the regularity of $\mu$ is equivalent to $ \operatorname{deg}(\mu)=\operatorname{deg}(\mu^h)$.

Diffusion models generate high-quality images, yet often violate the physical laws governing mirror reflections. Reflections often suffer from geometric aberrations, including positional offsets, directional misalignment, proportional imbalance, and structural distortion. These failures remain evident even in contemporary state-of-the-art generative systems. Existing methods itigate this problem through synthetic data scaling or auxiliary depth conditioning, yet their merely reliance on latent-space noise reconstruction losses as implicit supervision prevents direct enforcement of reflection-specific geometric and perceptual constraints. To bridge this gap, we present PhysReflect, a geometry and perception guided diffusion framework that decodes the predicted clean latent into pixel space at each training step and applies annealed supervision through two complementary differentiable objectives. The Geometric Loss enforces mirror-induced spatial consistency through sparse epipolar correspondence and dense boundary projection alignment, where a SAM2-based TwinTrack mechanism provides stable in-mirror localization for boundary-aware supervision. The Perceptual Loss preserves reflected appearance by combining Semantic Consistency Loss, which maintains reflected identity and appearance via DINOv2 features, and Lighting Consistency Loss, which regularizes depth, surface-normal, and illumination coherence under monocular geometry priors. Experiments on synthetic and real-world benchmarks show that PhysReflect outperforms prior mirror-reflection methods in geometric, perceptual, and physical-plausibility metrics, as well as qualitative visual results.

We determine the preshared entanglement required for spatially separated parties, restricted to local operations and classical communication, to attain globally optimal probabilistic two-copy purification of arbitrary bipartite pure states under depolarizing noise. In every local dimension $d\ge 2$, one shared maximally entangled qubit pair suffices: local controlled-SWAP operations exactly reproduce the globally optimal successful transformation. Conversely, any finite-dimensional preshared resource state that attains the same benchmark, even under a positive-partial-transpose relaxation, must have entanglement of formation at least one ebit. For pure resources with one ebit, or for two-qubit resources including mixed states, attaining the benchmark is possible only for states with exactly two nonzero Schmidt weights, both equal to $1/2$, namely those locally unitarily equivalent to a maximally entangled qubit pair. These findings provide a benchmark for evaluating the entanglement demands of noise management in quantum networks and modular quantum computers.

Let $G$ be a finite group. In this article, we develop some results of equivariant persistent homotopy theory for persistent $G$-spaces. We introduce the notions of $G$-stable and $G$-homotopy invariant distances and define the $G$-homotopy interleaving distance, an equivariant analogue $d^G_{HI}$ of the homotopy interleaving distance of Blumberg and Lesnick. We prove that this distance $d^G_{HI}$ is $G$-stable and $G$-homotopy invariant and establish its universality by showing that it dominates any such distance. We further prove a version of equivariant persistent Whitehead theorem relating interleavings of equivariant persistent homotopy groups to $G$-homotopy interleavings of persistent $G$-spaces. We also prove an equivariant persistent nerve lemma for equivariant good covers of persistent $G$-spaces. As a consequence, we obtain an equivariant weak law of large numbers for filtrations, providing a homotopy-theoretic consistency result for random filtrations equipped with finite group of symmetries.

Non-prehensile manipulation is practical for relocating large, heavy, or geometrically ungraspable objects. Yet, long-horizon pushing of arbitrarily-shaped 3D objects couples three problems: 1) where to push the object so as to approach the target pose, 2) whether each push is stable and reachable, 3) whether subsequent actions remain feasible. We present an object-centric pushing policy within a feedback-guided hierarchical framework. At the low level, a learning-based policy predicts contact actions from a pose- and scale-normalized point cloud, conditioned on a near single-step subgoal. A stability score is applied to evaluate the predicted contacts by a quasi-static sliding-versus-tipping analysis. At the high level, BIT$^*$ first searches for an object path, and the next several subgoals are checked by contact prediction and robot motion planning for future feasibility. Failed motion plans, as feedback, change the local path costs and trigger re-planning. During execution, only the first feasible action is executed. In simulation, we evaluate 22 objects in six different scenes, upon which we also conduct comprehensive ablation studies. Results demonstrate that our method outperforms baselines with a clear margin and can reliably achieve long-horizon object pushing tasks under different situations. We also report quantitative real-robot experiments with a Franka arm and qualitative demonstrations with a mobile manipulator for large and heavy objects, with directly zero-shot sim-to-real transfer.

Big data can be hosted on cloud and being shared distributedly through cloud services in an unprecedented volume, variety and velocity. This causes not only cloud network congestions and delayed cloud services but also increases in public cloud data-out charges. Client-side cloud cache alleviates these problems. Furthermore, cloud cache must be aware of nonuniform data-out costs when big data is stored in hybrid clouds built with different public cloud providers. Deploying i-Cloud approach as the core mechanism of cloud cache could save data-out cost up to 14.78% or 4,425 USD saved per annum based on our representative scenario, and delivered 17.24% byte-hit, 17.96% delay-saving and 29.33% cache hit outperforming LRU, GDSF and LFU-DA approaches. A main finding is that i-Cloud, learning uniform cost patterns, could perform well against nonuniform cost environment.

We investigate whether the plasma disturbances generated by a charged debris object can be used to infer its position without directly sampling the object or its immediate sheath. A one-dimensional open-boundary electrostatic particle-in-cell (PIC) framework is used to model a continuously flowing electron-ion plasma containing an initially uncharged debris object. The debris charges self-consistently and produces an ion-ion counter-streaming instability (IICSI), which is sustained and reaches a statistically stationary state. A matching simulation without debris provides a controlled background against which debris-induced changes in different plasma fields are identified. The precursor and wake are examined through their spatial extent, fluctuation power, and frequency–wavenumber spectra. Weak-form sparse regression is then used as a supporting tool to recover reduced fluid and kinetic residuals and to learn an empirical relation connecting the precursor and wake disturbance envelopes to the equilibrium plasma flow, debris charge, and distance from the source. The predictive capability is examined by successively treating each simulation as an independent case. For each evaluation, the model is constructed from the remaining simulations, while the excluded simulation is used only to infer the debris position from its plasma response. For the idealized cold-ion regime considered here, the procedure successfully localizes the debris in transonic and supersonic flows, whereas the subsonic cases remain unresolved. These results demonstrate that debris-induced plasma disturbances possess a learnable spatial structure and provide a proof-of-principle route toward future remote localization using more realistic multidimensional models; they do not constitute an operational debris-detection scheme.

Novel view synthesis from sparse observations is severely under-constrained. Although 3D Gaussian Splatting (3DGS) enables real-time rendering, it produces floaters, broken geometry, and washed-out backgrounds when trained with few views. We propose an alternating optimization framework that uses a pre-trained image diffusion model to generate geometrically consistent pseudo-views for additional 3DGS supervision. Generation is constrained by depth-conditioned ControlNet, IP-Adapter style transfer, LoRA scene adaptation, and img2img structural anchoring. We introduce Generative Active Pseudo-view Selection (GAPS) to balance reconstruction informativeness and generative reliability when choosing target views. Its annealing schedule shifts from conservative interpolation early in training to exploratory extrapolation later, gradually covering unobserved regions. A dual-criterion admission gate and uncertainty-weighted losses reject unreliable generations, while density-adaptive DropGaussian reduces overfitting in complex scenes. On LLFF with 3/6/9 views, our method improves average PSNR over vanilla 3DGS by 0.40/0.89/0.70 dB. On Mip-NeRF 360 with 12/24 views, the gains are 1.18/0.80 dB. SSIM improves and LPIPS decreases in every setting. Ablations show that active selection and density-adaptive regularization are both necessary; only the full method reduces LPIPS below the no-pseudo-view baseline on unbounded 360-degree scenes.

Multi-omics sequences contain complex biological patterns, yet deciphering their mechanisms for automated scientific discovery remains challenging. As large language models (LLMs) interpret these sequences, evaluating both predictions and scientific reasoning is critical. However, existing benchmarks for multi-omics sequence tasks rely on classification and regression metrics, neglecting whether models grasp the underlying biological evidence. We introduce OmicsBench, the first reasoning benchmark for multi-omics sequences, comprising 1,160 expert-validated questions across six tasks spanning DNA regulation, RNA processing, and protein function. OmicsBench requires traceable evidence chains, evaluated using instance-specific rubrics developed with domain experts. Evaluating 17 LLMs reveals an inverse relationship: while scientific LLMs outperform general-purpose LLMs in sequence classification accuracy, they fail to provide valid evidence to support their predictions. One plausible interpretation is shortcut learning: specialized models may rely on statistical patterns rather than the biological mechanisms needed for scientific discovery. Motivated by this finding, we introduce tool-augmented on-policy distillation (TA-OPD), a post-training method to align sequence prediction with evidence-grounded biological reasoning. Across five Qwen3.5 models spanning 0.8B to 27B parameters, TA-OPD consistently strengthens biological evidence grounding while improving predictive performance on most tasks. These gains persist across model scales, indicating that stronger sequence reasoning does not arise solely from increased model capacity, but can be improved through evidence-aware training. Together, OmicsBench and TA-OPD provide a framework for diagnosing reasoning failures in multi-omics LLMs and a path toward models whose predictions are better grounded in biologically meaningful evidence.

We study the correspondence between quasinormal modes (QNMs) and grey-body factors (GBFs) of higher-dimensional Reissner–Nordström black holes. Considering scalar, vector, and tensor gravitational perturbations, we systematically compute the QNMs and GBFs for each spacetime dimension and set of black-hole parameters to evaluate the analytical formula connecting them. QNMs are obtained by the continued fraction and the integration-through-midpoints methods depending on the singularity structure in the Frobenius series. We compare the GBFs determined by the correspondence using QNMs with those computed numerically. The correspondence performs well for scalar($+$), vector($+$), vector($-$), and tensor perturbations. For the scalar($-$) perturbation, the GBFs obtained by the correspondence deviate noticeably from the accurate GBFs for black holes with low to intermediate values of the charge in $D>6$. The correspondence recovers reasonable accuracy at higher charges, although the effective potential exhibits double peaks, contrary to the assumption underlying the Wentzel–Kramers–Brillouin (WKB)-based derivation of the correspondence. Whereas the accuracy decreases with increasing $D$ for other perturbation types, the correspondence for the scalar($-$) perturbation yields more accurate GBFs in $D=8$ than those in $D=7$. For near-extreme black holes, the correspondence offers a reliable approximation of GBFs for the vector($-$) perturbation, even though a double-peak potential occurs in $D>6$. These results suggest that the correspondence has a wider range of applicability than might be expected from the assumption of the WKB approach.

Voice anonymization aims to protect privacy by suppressing speaker identity while preserving linguistic content and prosody. However, residual speaker attributes in non-identity representations may still increase linkability and weaken privacy protection. To this end, this paper proposes a fine-tuning strategy with a pinhole loss for well-trained voice anonymization frameworks to further reduce residual speaker attributes. Inspired by the pinhole effect, the pinhole loss measures the linkability of anonymized utterances from the same source speaker. By minimizing this loss, linkability is reduced, thereby improving privacy protection. Experiments on multiple anonymization frameworks, pseudo-speaker generation methods, and datasets show improved privacy protection while maintaining utility. Audio samples can be found in https://anonymous.4open.science/r/Pinhole-loss-fine-tunning-4628.

Dynamic rope manipulation is highly sensitive to unknown object dynamics: the same robot motion can produce substantially different responses across ropes, while explicitly identifying the relevant physical properties is difficult. We present RopeFormer, a history-conditioned framework that uses prior task interaction as context for subsequent control. The policy retains cross-trial action-response history while keeping its weights fixed and requires no explicit online rope-parameter estimation. In matched simulation evaluations across sustained single-arm rotation, bimanual rotation, and transient whipping, retaining context improves subsequent control relative to resetting the same checkpoint, with the benefit varying across rope dynamics and observation settings. We further deploy the frozen policies on a Unitree H1-2 with previously unseen physical ropes. From T1 to T3, target-acquisition time decreases by 30.9% for Rope Swing and 33.9% for Rope Twirl, while mean Rope Whip target hits increase from 0.2 to 2.3 out of three. These results show that prior interaction can provide effective control context for dynamic deformable-object manipulation. Robot videos, code, and data are available at https://ropeformer.github.io/.

Human-object interaction (HOI) detection requires grounding an interacting human-object pair and recognizing the verb that links them, often under severe long-tail supervision. Recent methods improve accuracy with stronger detectors and vision-language priors, but many still stack heavy transformer encoders, intricate denoising schedules, or post-hoc semantic calibration on top of the detector. We present HOIBlender, an efficient HOI detector named after its core design principle: blending detector-grounded visual tokens, spatial subject-object reasoning, and BLIP-2 semantic priors inside one lightweight decoding pipeline. HOIBlender builds on an RF-DETR/LW-DETR-style foundation with a DINOv2 backbone and selects top-$K$ image-conditioned tokens directly from the multi-scale projector as subject and object candidates, removing the dedicated encoder stage retained by prior HOI methods. A dual-stage decoder first stabilizes human-object geometry and then performs verb and HOI classification through progressive BLIP-2 prior fusion, with classifier weights initialized from BLIP-2 text embeddings for long-tail categories. Grouped-query training further enriches optimization without increasing inference cost. Across three model scales (Nano, Small, 2XL), HOIBlender consistently outperforms SOV-STG-VLA and Hybrid-SOV-VLA on HICO-DET, reaching $44.49$ Default Full mAP in only $9$ training epochs while maintaining competitive latency and parameter budgets. These results show that lightweight detection, structured spatial-semantic decoding, and deeply integrated vision-language priors can be blended into a single efficient HOI pipeline.

In characteristic $p =3$, we establish a concrete relationship between the torsion rank of the elementary abelian group of global sections of quasi-elliptic surfaces over the projective line ${\mathbf P}^1$ and the $p$-rank of certain hyperelliptic curves, thereby explaining a mysterious relationship between these two invariants previously observed by I. Dolgachev. As an application, we determine the torsion ranks of the Mordell–Weil groups of a class of quasi-elliptic surfaces.

Open quantum systems offer a physical substrate for temporal information processing in quantum reservoir computing, yet the microscopic mechanisms linking their dynamics to computational performance remain unclear. Here we establish a microscopic response-to-performance framework that connects Liouvillian dynamics directly to task performance. We decompose observable Volterra weights as $W^{(q)}=BH^{(q)}$, separating internal input-history pathways from their readout visibility, and show that a dynamical response contributes to a task only if it is input-generated, readout-visible, target-correlated, and resolvable above the regularization scale. This framework reveals how Liouvillian modes govern the retention and propagation of task-relevant information, while symmetries determine which components remain visible to the readout. It further establishes how distinct covariance modes encode different temporal information structures and how system parameters and processing protocols can reshape their contributions to improve QRC performance. Thermalization drives the reservoir from processing that retains input history to a response dominated by the most recent input before regularization suppresses the residual information. We finally show that weak measurement can activate task-relevant covariance modes by lifting symmetry-imposed channel equivalence, whereas delayed feedback creates controlled return pathways for historical information. These results establish a microscopic connection between open-system dynamics and computational performance, providing principles for engineering task-relevant information flow in quantum reservoirs.

Dark matter substructure properties such as their mass function and spatial distribution depend on the nature of dark matter and are strong tests of the cosmological model. Like luminous matter, these dark matter substructures cause gravitational lensing affecting observables such as time delays. Given the millisecond-level timing precision achievable with current and future gravitational-wave (GW) detectors, gravitationally lensed GWs provide a powerful probe of dark matter substructure, especially, at the lower end of the mass function. In contrast, the optical (or electromagnetic) observations can provide a precision of a few hours and are thus, not sensitive to deflections from the less massive subhalos. In this work, we investigate the impact of realistic population of DM subhalos ($10^6-10^9 M_\odot$) on the lensing time delays for two of the typical strong lens configurations, a fold and a cusp, seen in quadruply lensed sources for a galaxy-scale lens. We find that the subhalos with NFW density profiles cause perturbations of the order of few hours to the lensing time delays between the macro-lensed images produced by the main lensing galaxy. These time delay "anomalies", if not accounted for, may affect the results of strong lens searches conducted on the GW data by the LIGO–Virgo–Kagra collaboration. Lastly, for the 200 realisations analysed per fold and cusp lens systems, we find that the NFW subhalos produced no additional images of their own called milli-images. Statistical studies are needed to better determine the expected impact on the lensed GW time delays and (non-)detection of milli-images.

Vision foundation models for Earth observation (EO) are commonly evaluated on clean benchmarks, but operational EO products can already contain spatial, radiometric, alignment, noise, and harmonization defects before reaching the model. Generic image corruptions and broad domain shifts do not isolate these product-level failures. We introduce RSPDBench, a physically grounded remote-sensing-product degradation benchmark that evaluates seven foundation-model entries and two supervised baselines on five EO datasets under primitive degradations and compound product chains. Its degradation parameters follow published Sentinel-2 and Landsat~8 specifications and are checked against Sentinel-2's paired $10\,\mathrm{m}$ and $20\,\mathrm{m}$ near-infrared bands. We evaluate each model under its own native protocol and measure robustness as the drop from its own clean baseline. Our analysis reveals that degradation sensitivity is strongly structured: (i) the resolution-conditioned encoder is the most robust to resolution loss yet the most sensitive to radiometric gain on satellite products, (ii) the same physical defect can hurt one model while helping another, and (iii) model rankings under resolution loss and radiometric gain are largely reversed. Compound chains add up to $38$ percentage points of drop beyond their strongest component. On m-BigEarthNet, retraining the probe on degraded views removes about 75% of the noise and harmonization losses, but resolution loss still costs about 20% of the clean score. EO robustness must therefore be measured against the structured defects of remote-sensing products, not by clean accuracy or generic perturbations alone. Code: https://github.com/Tanjim131/RSPDBench

Starting from the unitary evolution of a closed quantum system, we rigorously demonstrate that the projection of the global wavefunction onto an arbitrary local subsystem is governed by an exact, time-dependent non-Hermitian Schrödinger equation. Crucially, the derivation does not rely on conventional approximations such as the Born approximation, the Markov approximation, or the wide-band limit. The central quantity is the instantaneous self-energy, a time-dependent and generally non-Hermitian operator that encodes environmental backaction and, together with the subsystem Hamiltonian, forms the exact time-local generator of the projected dynamics. By benchmarking the conventional non-Hermitian approximation against this exact framework, we systematically expose its limitations. These results provide a rigorous microscopic foundation for the emergence of effective non-Hermitian dynamics in quantum systems.