arXiv:2607.11915v1 Announce Type: new
Abstract: Plasma diagnostic models for tokamak fusion devices are almost universally evaluated on clean, complete sensor data. In practice, fusion diagnostics fail regularly: acquisition systems start late, individual sensors die, and signal dropouts cluster precisely when a plasma disruption is approaching. We present the first systematic robustness benchmark for plasma diagnostic ML using the TokaMark dataset of 11,573 MAST shots, evaluating XGBoost, LSTM, Transformer, and the TokaMark CNN baseline across six physically-grounded failure scenarios and three imputation strategies. We introduce the Robustness Score (RS) for standardized cross-architecture comparison. Our central finding is that disruption-proximate sensor failure (corruption injected in the final window timesteps) collapses sequence model performance (LSTM +212% NRMSE) while a statistical feature model remains comparatively stable (XGBoost +37%). Forward-fill imputation eliminates nearly all degradation from random dropout for sequence models (LSTM +57% to ~0%), but offers little help when the end of the window is corrupted. Shot-level alarm evaluation using ground-truth disruption timestamps reveals that LSTM alarm detection collapses to TPR=0.00 under proximate sensor failure, while mean-fill imputation recovers it to TPR=1.00, a reversal of the pattern observed in NRMSE. Plasma current emerges as the single most critical diagnostic across all architectures (+73% to +140% upon removal). Code, data, and trained checkpoints are available at https://github.com/Neerav-Gupta/tokamark-robustness.
Science Journals
arXiv:2607.11916v1 Announce Type: new
Abstract: The integration of Large Language Models (LLMs) with evolutionary computation has emerged as a powerful paradigm for automated heuristic design in combinatorial optimization. However, existing approaches suffer from mode collapse, converging to homogeneous populations that lack semantic diversity and fail to explore the full algorithmic space. We propose Quality-Diversity Evolution (QDEvo), a multi-objective framework that integrates Quality-Diversity optimization with LLM-driven heuristic search, maintaining an unbounded archive of semantically diverse algorithms using pre-trained code embeddings and incorporating hierarchical self-reflection to guide the evolutionary process. Extensive experiments across standard benchmarks and real-world industrial applications demonstrate that QDEvo significantly outperforms state-of-the-art methods in both Hypervolume and Inverted Generational Distance metrics. Our framework enables the discovery of heuristics that are simultaneously high-performing, computationally efficient, and semantically diverse, providing practitioners with a rich portfolio of solutions for complex optimization problems.
arXiv:2607.11940v1 Announce Type: new
Abstract: As the scale of large pre-trained models continues to grow, fine-tuning them under limited memory budgets has become increasingly challenging. Low-Rank Adaptation (LoRA), currently one of the most widely adopted parameter-efficient fine-tuning (PEFT) methods, mitigates this challenge by optimizing only low-rank adaptation matrices, thereby greatly reducing the number of trainable parameters. With the parameter overhead substantially reduced, the activations retained for backpropagation have emerged as the primary remaining memory bottleneck during LoRA fine-tuning. To address this, we propose CARE-LoRA, a data-aware Compressed Activation REconstruction framework. By exploiting the inherent projection structure of LoRA, CARE-LoRA replaces the full input activation with the low-rank compressed activation naturally produced by the LoRA branch. It further computes a lightweight reconstruction matrix during the forward pass with negligible additional computation cost, which is used during backpropagation to reconstruct the gradient signal, thereby keeping LoRA matrices fully trainable. Extensive experiments across diverse models and downstream tasks demonstrate that, while substantially reducing the overall memory footprint, CARE-LoRA achieves competitive or even superior performance compared with standard LoRA and representative LoRA variants. Our code is publicly available at https://github.com/fishandyu/CARE-LoRA .
arXiv:2607.11964v1 Announce Type: new
Abstract: Autonomous driving requires long-horizon closedloop decision making in dynamic traffic environments. Latent world models offer an effective framework for this problem by enabling imagination-based decision making in compact latent spaces. However, multi-source observations contain controlirrelevant redundancy, whereas reliable driving decisions rely on risk-relevant relations, future dynamics, and continuous action adjustments. This mismatch makes observation reconstruction and absolute action modeling suboptimal for learning decisionrelevant latent dynamics. We propose LIDAR-AD, a decoderfree Latent-Interaction Dreamer with Action-Residual Chains for autonomous driving. LIDAR-AD replaces observation reconstruction with redundancy-reduced latent alignment, encouraging compact representations of risk-relevant relations in multi-source driving inputs. It further models vehicle control as residual action updates and uses residual-action sequence contrastive learning to align multi-step residual-driven rollouts with future latent states. A deterministic analysis shows that the latent-tanh residual parameterization preserves interior action reachability while representing smooth long-horizon control as compact local updates. Together, these designs improve risk-aware state abstraction, continuous-control modeling, and long-horizon dynamics prediction. Extensive experiments across diverse simulated driving scenarios demonstrate that LIDAR-AD consistently outperforms world-model baselines, achieving the highest reward and the best success rate among learning-based methods. Evaluations on nuPlan-derived log-reconstructed scenarios further demonstrate the transferability of LIDAR-AD under real-world traffic layouts.
arXiv:2607.12329v1 Announce Type: new
Abstract: In human dialogue, we achieve smooth communication by expressing nonverbal cues such as eye contact, nodding, and facial expressions with precise timing. It is expected for conversational avatars to express these cues appropriately to realize natural and human-like interactions. This study focuses on nodding, which is crucial for demonstrating active listening and encouraging further user utterances. We propose a model that predicts both timing and kinematic parameters representing the motion features of listener nodding in real time. The proposed model consists of a timing prediction module and a kinematic parameter prediction module. Each implements a dyadic attention network over the speaker and listener channels based on the technique of Voice Activity Projection (VAP). Unlike conventional models, this approach enables real-time prediction of kinematic parameters based on the specific context of the dialogue rather than just predicting the timing. Furthermore, we demonstrate the effectiveness of fine-tuning the kinematic parameter prediction module initialized from the trained timing prediction module. The proposed model is lightweight and capable of real-time operation, and it has been integrated into an avatar dialogue system. Subjective evaluation experiments shows that our proposed method significantly outperforms both a baseline with stochastic timing and another with fixed-motion nodding. The code and trained models are available at https://github.com/MaAI-Kyoto/MaAI.
arXiv:2511.11498v2 Announce Type: replace
Abstract: We consider the problems of \emph{learning} and \emph{testing} real-valued convex functions over Gaussian space. Despite the extensive study of function convexity across mathematics, statistics, and computer science, its learnability and testability have largely been examined only in discrete or restricted settings -- typically with respect to the Hamming distance, which is ill-suited for real-valued functions.
In contrast, we study these problems in high dimensions under the standard Gaussian measure, assuming sample access to the function and a mild smoothness condition, namely Lipschitzness. A smoothness assumption is natural and, in fact, necessary even in one dimension: without it, convexity cannot be inferred from finitely many samples. As our main results, we give:
- Learning Convex Functions: An agnostic proper learning algorithm for Lipschitz convex functions that achieves error $\varepsilon$ using $n^{O(1/\varepsilon^2)}$ samples, together with a complementary lower bound of $n^{\mathrm{poly}(1/\varepsilon)}$ samples in the \emph{correlational statistical query (CSQ)} model.
- Testing Convex Functions: A tolerant (two-sided) tester for convexity of Lipschitz functions with the same sample complexity (as a corollary of our learning result), and a one-sided tester (which never rejects convex functions) using $O(\sqrt{n}/\varepsilon)^n$ samples.
arXiv:2511.17239v2 Announce Type: replace
Abstract: This paper studies three structured approximation problems: (1) Recovering the range of a Fourier matrix from a single observation, (2) Recovering a corrupted low-rank Toeplitz/Hankel matrix, and (3) Recovering a finite exponential sum from noisy samples. All three problems are computationally challenging because their structural constraints are difficult to enforce directly. We show that all three tasks can be solved efficiently and optimally by applying the Gradient-MUSIC algorithm for spectral estimation. To provide an example, for a rank-$r$ Toeplitz matrix $T\in {\mathbb C}^{n\times n}$ that satisfies a regularity assumption and is corrupted by an arbitrary $E\in {\mathbb C}^{n\times n}$ such that $\|E\|_2\leq \alpha n$, our algorithm outputs a Toeplitz matrix $T_\sharp$ of rank exactly $r$ such that $\|T-T_\sharp\|_2 \leq C \|E\|_2$, where $C,\alpha>0$ are absolute constants. This performance guarantee is minimax optimal in $n$, $r$, and $\|E\|_2$. For the other two structured approximation problems, we also provide algorithms that are minimax optimal in the number of samples, rank/sparsity, and noise level. At the heart of this paper is a quantitative transference principle which shows how to convert computational methods and theory for spectral estimation into corresponding methods and theory for the other three problems.
arXiv:2607.11673v2 Announce Type: replace
Abstract: We present ABot-3DWorld 0, a universal multimodal 3D world model that turns text, image, and video inputs into high-fidelity, explorable 3D worlds. At the heart of our framework is a unified Spatial Generative Primitive (SGP), a compact tuple of a high-quality panorama and a spatial point cloud that delivers an efficient description of any 3D space. Multimodal inputs are first lifted into this primitive; a 3D-consistent panoramic video generator then explores the primitive along a planned trajectory; finally, our panoramic video reconstruction engine converts the generated video into a clean, photorealistic 3D Gaussian Splatting (3DGS) world. This pipeline covers two regimes: rich inputs (multi-view sets, casual video) are lifted into the SGP through a geometry-rigorous recovery that mirrors the observed scene, while a single image or sentence is completed generatively into a creative world. The result is one low-barrier engine for general 3D content creation that further anchors generated worlds to geographic points of interest, enabling map-native spatial exploration at consumer scale. Experiments show that ABot-3DWorld 0 sets the state of the art among open-source methods and demonstrates stronger scene fidelity than Marble under rich multimodal inputs.
arXiv:2607.12153v1 Announce Type: new
Abstract: Splitting methods are among the most classical and fundamental tools for the simulation of quantum dynamics, and their importance has grown further with the rise of quantum computing. In this work, we analyze the Schr\"odinger equation with Yukawa potential, a physically relevant and widely used model potential. It may be viewed as a Coulomb interaction with exponential decay at spatial infinity, preserving the Coulomb singularity at the origin while removing the long-range Coulomb tail. We prove that the operator splitting for this unbounded Hamiltonian achieves a global $1/4$-order convergence rate in the time step for many-body Yukawa interactions, with explicit polynomial dependence on the number of particles. The result holds for all initial wavefunctions in $H^2(\mathbb R^{3N})$, the natural domain of the Hamiltonian, and our numerical experiments are consistent with the theoretical estimates. To identify the sharp obstruction behind this rate, we prove a short-time lower bound in the one-body setting of order $t^{5/4}$ for the one-step error, which rules out any uniform global estimate of order better than $1/4$ in general. This agreement with the optimal $1/4$ rate in the Coulomb case is particularly interesting, as Yukawa potential is short-ranged compared to Coulomb potential. For the many-body upper bound, one of the new technical ingredients is the explicit polynomial-in-system-size Sobolev estimates of many-body Yukawa systems. These estimates are crucial for obtaining fully a priori bounds that depend only on the norms of the initial states, rather than on the solution at time $t$. For the one-body lower bound, we leverage a new analysis argument based on Fourier analysis and Kato smoothing.
arXiv:2607.11987v1 Announce Type: new
Abstract: Perineural invasion (PNI) is associated with poor postoperative outcomes in intrahepatic cholangiocarcinoma, but it is confirmed by surgical pathology. Existing preoperative imaging models often rely on radiologist-defined variables, contrast-enhanced imaging, or manual annotations. We propose an anatomy-privileged teacher--student framework for patient-level PNI prediction from T2-weighted MRI. During training, the teacher uses MRI with tumor and liver masks to learn dense token routing, and the student distills this guidance to retain and aggregate informative tokens under a fixed budget. Anatomical supervision is restricted to training, and the deployed model does not require masks at inference. In 155 patients, the proposed method achieved the highest mean AUROC of 0.750 among matched MRI-only baselines evaluated under the same protocol, with 1.43 GFLOPs and 8.02 ms per case on a Jetson Orin Nano Super Developer Kit.
arXiv:2607.12167v1 Announce Type: new
Abstract: Astronomical observations allow us to better our understanding of the universe. As we observe smaller and more distance features, we run into the diffraction limit of our observation system. This limit is a function of the wavelength observed and the size of the primary aperture used. We can synthesize a larger primary aperture by implementing interferometry. Long baseline optical interferometry would lead to significant improvements in astronomical imaging resolution. Building optical interferometers with free space baselines becomes significantly more difficult as the baseline increases. A promising alternative is to use optical fiber to connect telescopes. Fiber-based interferometers are much more susceptible to phase noise than their free-space counterparts due to inhomogeneities in the fiber medium. This leads to significant degradation of interferometric signals used for astronomical measurements. We implement phase stabilization techniques used in quantum communications to stabilize an optical interferometer with a 170km pseudo-baseline. We use a quantum optimal measurement technique with this interferometer to resolve the extent of a source four times smaller than the diffraction limit of the system within 1.5% error. These results bring the potential for a full-scale on-sky long baseline interferometer significantly closer. A 350km baseline optical interferometer at 1550nm would allow us to resolve sub microarcsecond features in the universe.
arXiv:2607.12180v1 Announce Type: new
Abstract: An AI teammate's design properties (personality, communication style, when it speaks) can shape a team's trust, coordination, and decisions. Studying this rigorously demands infrastructure no existing tool provides: reproducible configuration of an AI teammate embedded in instrumented, real-time collaboration sustained over time. We present the Team Research and AI Integration Lab (TRAIL), a web platform that makes the AI teammate a configurable, reproducible design object, pairing a Big Five persona with a selective-participation message pipeline, dual memory, chained longitudinal experiments, and export-ready analytics. In a real six-session classroom deployment (about 51 students), TRAIL sustained longitudinal chaining, held the AI to a stable minority of the conversation, and enabled export-driven AI-human text-similarity analysis. A single blind persona change produced a design-consistent double dissociation: a cognitive-scaffolding agent drew stronger contribution ratings and closer linguistic alignment; a socially-supportive agent, a warmer team climate and lower over-reliance.
arXiv:2502.15131v5 Announce Type: replace-cross
Abstract: We study the fundamental problem of calibrating a linear binary classifier of the form $\sigma(\hat{w}^\top x)$, where the feature vector $x$ is Gaussian, $\sigma$ is a link function, and $\hat{w}$ is an estimator of the true linear weight $w^\star$. By interpolating with a noninformative $\textit{chance classifier}$, we construct a well-calibrated predictor whose interpolation weight depends on the angle $\angle(\hat{w}, w_\star)$ between the estimator $\hat{w}$ and the true linear weight $w_\star$. We establish that this angular calibration approach is provably well-calibrated in a high-dimensional regime where the number of samples and features both diverge, at a comparable rate. The angle $\angle(\hat{w}, w_\star)$ can be consistently estimated. Furthermore, the resulting predictor is uniquely $\textit{Bregman-optimal}$, minimizing the Bregman divergence to the true label distribution within a suitable class of calibrated predictors. Our work is the first to provide a calibration strategy that satisfies both calibration and optimality properties provably in high dimensions. Additionally, we identify conditions under which a classical Platt-scaling predictor converges to our Bregman-optimal calibrated solution. Thus, Platt-scaling also inherits these desirable properties provably in high dimensions.
arXiv:2607.03999v2 Announce Type: replace-cross
Abstract: Estimating heterogeneous treatment effects (CATE) requires simultaneously detecting effect modification and quantifying estimation uncertainty. Existing tree-based methods make an uneasy trade-off: significance-based approaches (Radcliffe and Surry 2011) identify subgroup interactions directly but lack valid inference; honest causal trees (Athey and Imbens 2016) deliver nominal confidence interval coverage but use outcome-agnostic splitting criteria that sacrifice interaction sensitivity. We introduce a hybrid algorithm that fuses significance-based splitting with honest sample-splitting and cross-validation. Our splitting criterion uses the squared $t$-statistic for the treatment $\times$ side interaction ($t^2$), which is shown to be directly aligned with the honest $\text{EMSE}_\tau$ criterion when the interaction is strong. Post-hoc honest cross-validation selects the cost-complexity penalty, giving a single principled estimator with nominal CI coverage at the leaf level. For forests, we retain bootstrap count vectors to enable an infinitesimal jackknife (IJ) variance estimate of Monte-Carlo convergence rather than formal pointwise inference. On the three synthetic designs from (Athey and Imbens 2016) the single tree achieves approximately 90% leaf-average CI coverage at the 90% nominal level across all three designs (200 replications each); on the Criteo, Hillstrom and Starbucks uplift datasets we match Qini coefficient performance of S-, T-learner and GRF baselines. An open-source Python package with reproducible seeds, sklearn-compatible API, and full test coverage accompanies this work (https://codeberg.org/hadjipantelis/rattus).
arXiv:2602.12823v5 Announce Type: replace-cross
Abstract: We present a technique for measuring ion temperature using cavity-based electromagnetically induced transparency (EIT) applicable for cavity QED systems. This method enables efficient extraction of the ion's phonon occupation number following sub-Doppler cooling close to the motional ground state. The proposed method requires operation in the resolved-sideband regime, where individual motional states can be selectively addressed for all relevant transitions either by selecting appropriate energy levels for the three-level system or by employing strong confinement with high secular frequencies ($\sim 10 MHz$). It relies on monitoring the cavity probe transmission while scanning the probe laser frequency to establish cavity-induced EIT using a control beam, thereby significantly simplifying the measurement procedure. We establish a theoretical model that demonstrates the influence of the thermal state of the trapped ion vis-\`a-vis the EIT linewidth measured. We show through numerical simulations how the cavity-induced EIT transmission may be used as a thermometry tool to deduce the ion temperature as well as its motional state in the sub-Doppler cooling regime, even for systems that are in the weak coupling regime.
arXiv:2607.12148v1 Announce Type: new
Abstract: We report on the design and evaluation of MindReader, a tool that helps a user replace her password when she is required to do so. Left to their own devices, users tend to replace their previous passwords with predictable variations of the original ones. MindReader leverages LLMs to suggest password variations that are chosen to be easy for the user to remember but harder for an attacker to predict. To do this, MindReader infers the meaning behind original password components and then suggests semantically related (yet syntactically unrelated) components for the new password. In a user study, passwords created using MindReader were more secure than both replacement passwords created without using MindReader and original passwords. In particular, MindReader replacement passwords were harder to guess in an online attack than alternative replacement passwords even by an attacker with knowledge of the original password and full knowledge of the tool implementation. Passwords created with MindReader were also comparably memorable to alternative replacement passwords and original passwords, as measured by the ability of users to successfully log in a week after creating their password.
arXiv:2603.22130v2 Announce Type: replace-cross
Abstract: We investigate how non-Markovian mechanical dissipation affects exceptional points in linearized optomechanical systems with red-sideband drive. For a chosen non-Ohmic mechanical bath, we derive analytical conditions for the memory-renormalized exceptional point by employing a pseudomode mapping, thereby demonstrating that structured environments displace the mode coalescence away from the Markovian prediction. Crucially, we reveal that failing to account for this memory-induced shift suppresses the divergent Petermann factor by orders of magnitude, showing that accurate bath modeling is essential for the successful operation of exceptional-point-based devices whenever reservoir-induced memory is non-negligible. We finally show that non-Markovianity modifies the cavity reflection spectrum, manifesting as a shallower optomechanically-induced-transparency dip, providing therefore an experimentally-accessible signature of structured mechanical environments.
arXiv:2604.19030v2 Announce Type: replace-cross
Abstract: Knitted fabrics exemplify a broad class of architected materials capable of large deformations, enabling shape morphing, mechanical biocompatibility, and embedded multifunctionality without material damage. Although geometric nonlinearity has been intuitively utilized in their design, a quantitative description of stitch-resolved deformation and its temporal evolution remains lacking. Here, we introduce a geometric quantification framework that reconstructs smooth yarn centerlines and fabric surfaces from sparse yarn-level representations and extracts interpretable descriptors across dimensions. Applied to representative knitted structures, this framework resolves how global deformation is distributed among stitch reorientation, loop bending, surface bending, and dilation. Moreover, it reveals how regions of large geometric variation emerge, persist, and redistribute over time. Rather than directly measuring stress, these geometric descriptors define a unified geometric state space for comparing knitted structures and identifying candidate regions of mechanical localization. The framework provides a quantitative language for nonlinear deformation in knits and establishes a geometry-based representation that can be coupled to constitutive models, experimental measurements, and graph-based inverse-design workflows.
arXiv:2607.10526v2 Announce Type: replace
Abstract: Stateful personal agents increasingly maintain long-term user profiles, episodic memories, and reusable skills. This persistence turns conversational sycophancy into a state-writing failure: accepted user-centric claims can be committed as lasting preferences, background facts, or workflows and later reused after the original conversation is gone. We call this persistent sycophancy and introduce the Personal Agent Sycophancy Benchmark (PASB), a 1,600-task benchmark that traces whether a conversational claim is accepted, written into durable agent state, and reused in a later neutral query. Unlike prior benchmarks that provide pre-written memories, PASB evaluates real agents (Hermes-Agent and OpenClaw) that decide what to store. It isolates the write process by combining four scenario framings with four temporal delivery patterns and separating a five-turn persist stage from a cleared three-turn query stage, ensuring downstream effects arise only from durable state. Across twelve models, the commit boundary is the key inflection point: downstream failure increases from 45.0% in session-only episodes to 71.9% after commitment, a consistent increase of 27.0 percentage points. Committed claims exhibit three write-time patterns: status promotion, attribution removal, and scope broadening. These patterns become stronger under memory-like or procedural framing, repeated reinforcement, and even across domain boundaries. These results show that agent sycophancy is fundamentally a state-writing governance problem. Once user content is committed to durable memory, safety must govern what agents write, not only what they say. PASB identifies the write-time controls needed to gate risky commits while preserving the source, role, and scope of stored content beyond response-level mitigations.
arXiv:2603.05995v2 Announce Type: replace
Abstract: Off-road autonomous driving poses significant challenges such as navigating unmapped, variable terrain with uncertain and diverse dynamics. Addressing these challenges requires effective long-horizon planning and adaptable control. Reinforcement Learning (RL) offers a promising solution by learning control policies directly from interaction. However, because off-road driving is a long-horizon task with low-signal rewards, standard RL methods are challenging to apply in this setting. We introduce TADPO, a novel policy gradient formulation that extends Proximal Policy Optimization (PPO), leveraging off-policy trajectories for teacher guidance and on-policy trajectories for student exploration. Building on this, we develop a vision-based, end-to-end RL system for high-speed off-road driving, capable of navigating extreme slopes and obstacle-rich terrain. We demonstrate our performance in simulation and, importantly, zero-shot sim-to-real transfer on a full-scale off-road vehicle. To our knowledge, this work represents the first deployment of RL-based policies on a full-scale off-road platform.
arXiv:2607.09967v2 Announce Type: replace
Abstract: Many neural networks operations have a multiplicative nature rather than additive: halving or doubling a norm are analogous relatively but require unequal optimization distances when taking linear steps. Adaptive optimizers such as Adam normalize updates per coordinate, but update steps remain additive; weights with very different magnitudes receive similarly sized absolute changes, producing very different relative perturbations. We introduce \textbf{\method} (\textbf{\methodshort}), a weight reparameterization for neural networks that combines a sign-aware symmetric-exponential pathway with an identity-like linear pathway. The symmetric-exponential pathway is near-linear for small raw weights but increasingly curved at larger magnitudes. Additive updates in logarithmic space map to magnitude-proportional changes in effective weight space. The linear pathway provides a direct route through the transform that we hypothesize stabilizes optimization, while learnable scale, curvature, and offset parameters control balance between pathways and the curvature of the exponential pathway. These components create a curved parameter-space geometry that empirically improves speed of loss descent over standard linear parameterization. We also identify a useful \emph{mismatched initialization}: raw weights are chosen so a symmetric version of the transform matches Xavier statistics, but training uses an asymmetric forward transform that leaves positive weights at full strength while making negative weights smaller in magnitude; in small-model ablations, this improves early optimization and may act as a form of symmetry breaking. We train transformers on OpenWebText over nine width$\times$depth configurations, \methodshort reaches matched validation loss in 1.32--1.49$\times$ fewer training steps, with the largest widths seeing the biggest gains.
arXiv:2607.12281v1 Announce Type: new
Abstract: Transformer-style architectures are increasingly adopted for industrial recommendation systems, yet they inherit a design premise misaligned with the task: generative models rely on per-token autoregressive prediction, which justifies maintaining large intermediate tensors that scale with sequence length. In contrast, recommendation systems produce a single set of relevance scores for each <user, item> pair without token-level supervision. Leveraging this observation, we propose SlimPer, which reformulates personalized ranking as iterative refinement of a compact, unified <user, item> knowledge base. At each layer, the model selectively queries raw multi-modal user-side tokens, computes explicit relevance matching scores, and refines the knowledge base, all in O(N) per-layer cost with a fixed-size intermediate representation. As a result, model depth is decoupled from user history length, enabling deeper relevance understanding without proportional growth in compute or memory; request-only optimization further trims memory by sharing a single copy of user-side tokens across all candidate items. SlimPer unifies sparse, dense, and sequence features within a single backbone and provides inherent interpretability through its attention mechanism. Deployed on Instagram Reels and Feed, SlimPer yields measurable improvements in user engagement while streamlining the overall system and enabling effective modeling of 10k+ fine-grained user history events.
arXiv:2602.12956v2 Announce Type: replace-cross
Abstract: We study the molecular beam epitaxy of self-assembled Al$\mathrm{_{1-x}}$Sc$\mathrm{_{x}}$N nanowires on conductive TiN layers and demonstrate their application in piezoelectric nanogenerators. Wurtzite Al$\mathrm{_{1-x}}$Sc$\mathrm{_{x}}$N nanowires with uniform Sc incorporation are grown across a wide composition range (0<x<0.35). At substrate temperatures below 700 $^\circ{}$C, these nanowires exhibit an inversely tapered morphology, whereas higher temperatures favor the nucleation of additional branches due to a phase separation of Al$\mathrm{_{1-x}}$Sc$\mathrm{_{x}}$N into wurtzite AlN and rock-salt ScN. Phase-pure Al$\mathrm{_{1-x}}$Sc$\mathrm{_{x}}$N nanowires are integrated into vertical nanogenerators, where the metallic TiN substrate serves as bottom electrode. The fabricated polymer-nanowire composite devices achieve effective piezoelectric charge coefficients of up to 8.5 pC N$^{-1}$ at x=0.32, thus exceeding the piezoelectric response of bulk AlN by nearly a factor of two. Although the charge response remains lower compared to Al$\mathrm{_{1-x}}$Sc$\mathrm{_{x}}$N thin films, the reduced effective dielectric permittivity of the nanowire-polymer composites compensates the reduction in piezoelectric charge coefficient, eventually yielding a higher voltage response and comparable energy harvesting efficiency. Finally, effective medium modeling reveals that the device architecture is the primary factor limiting performance, providing general design principles for highly efficient nanowire-based piezoelectric energy harvesters.
arXiv:2607.11959v1 Announce Type: new
Abstract: Greenhouse reinforcement learning can test climate-control ideas at a speed and scale that is difficult to achieve with crop experiments alone. For smart-greenhouse control, however, a single simulator return is not enough: a grower or control engineer also needs to know when the policy heats, enriches CO2, vents, manages humidity, deploys screens, or uses lamps.We propose a reproducible calibration-first reward audit framework that keeps named greenhouse-control reward components comparable across simulator training, facility-adapted rollouts, logged Autonomous Greenhouse Challenge records, and actuator-rule distillation. In GreenLight-Gym, the framework decomposes the scalar reward into conditional temperature, CO2, humidity and vapor-pressure-deficit, screen, and actuation-proxy terms; adapts GreenLight to the second Autonomous Greenhouse Challenge logged climate traces; and scores the same components on logged greenhouse data.
arXiv:2607.12612v1 Announce Type: new
Abstract: BERT models have revolutionised Natural Language Processing (NLP) through their ability to process unstructured text across diverse domains. However, developing high-quality BERT models for non-English languages remains challenging due to limited annotated data and high computational demands. Translating non-English data into English and fine-tuning existing English BERT models offers a resource-efficient alternative, yet few studies have structurally compared translation-based fine-tuning with native-language BERT performance across tasks and languages. This study provides such a comparison, evaluating the feasibility of translation-based fine-tuning across six NLP tasks: Sentiment Analysis, Hate Speech Detection, Question Answering, Named Entity Recognition, Part-of-Speech Tagging, and Natural Language Inference, using datasets translated from Bulgarian, Chinese, Dutch, Italian, and Russian. Across all settings, the translation-based approach was comparable or superior in 53.3 percent of cases. Gains were most frequent in Question Answering, Part-of-Speech Tagging, and Natural Language Inference, while performance declines were common in Named Entity Recognition and Hate Speech Detection. The results show that translation-based fine-tuning is most effective for tasks relying on syntactic or structural patterns and for languages typologically close to English, such as Dutch, but less effective for token-level or culturally nuanced tasks, particularly in Chinese. Overall, this study demonstrates that translation-based fine-tuning offers a scalable, resource-efficient, and empirically validated path for extending NLP to low-resource languages while advancing linguistic inclusivity and sustainability in artificial intelligence.