Deep reinforcement learning (DRL) agents excel in solving complex decision-making tasks across various domains. However, they often require a substantial number of training steps and a vast experience replay buffer, leading to significant computational and resource demands. To address these challenges, we introduce a novel theoretical result that leverages the Neyman-Rubin potential outcomes framework into DRL. Unlike most methods that focus on bounding the counterfactual loss, we establish a causal bound on the factual loss, which is analogous to the on-policy loss in DRL. This bound is computed by storing past value network outputs in the experience replay buffer, effectively utilizing data that is usually discarded. Extensive experiments across the Atari 2600 and MuJoCo domains on various agents, such as DQN and SAC, achieve up to 2,427% higher reward ratio, outperforming the same agents without our proposed term, and reducing the experience replay buffer size by up to 96%, significantly improving sample efficiency at negligible cost.
Large Language Model (LLM) agents are commonly tuned with supervised finetuning on ReAct-style expert trajectories or preference optimization over pairwise rollouts. Most of these methods focus on imitating specific expert behaviors or promoting chosen reasoning thoughts and actions over rejected ones. However, without reasoning and comparing over alternatives actions, LLM agents finetuned with these methods may over-commit towards seemingly plausible but suboptimal actions due to limited action space exploration. To address this, in this paper we propose Self-taught ActioN Deliberation (SAND) framework, enabling LLM agents to explicitly deliberate over candidate actions before committing to one. To tackle the challenges of when and what to deliberate given large action space and step-level action evaluation, we incorporate self-consistency action sampling and execution-guided action critique to help synthesize step-wise action deliberation thoughts using the base model of the LLM agent. In an iterative manner, the deliberation trajectories are then used to finetune the LLM agent itself. Evaluating on two representative interactive agent tasks, SAND achieves an average 20% improvement over initial supervised finetuning and also outperforms state-of-the-art agent tuning approaches.
Off-road autonomous navigation is a challenging task as it is mainly dependent on the accuracy of the motion model. Motion model performances are limited by their ability to predict the interaction between the terrain and the UGV, which an onboard sensor can not directly measure. In this work, we propose using the DRIVE protocol to standardize the collection of data for system identification and characterization of the slip state space. We validated this protocol by acquiring a dataset with two platforms (from 75 kg to 470 kg) on six terrains (i.e., asphalt, grass, gravel, ice, mud, sand) for a total of 4.9 hours and 14.7 km. Using this data, we evaluate the DRIVE protocol's ability to explore the velocity command space and identify the reachable velocities for terrain-robot interactions. We investigated the transfer function between the command velocity space and the resulting steady-state slip for an SSMR. An unpredictability metric is proposed to estimate command uncertainty and help assess risk likelihood and severity in deployment. Finally, we share our lessons learned on running system identification on large UGV to help the community.
WonderPlay is a novel framework integrating physics simulation with video generation for generating action-conditioned dynamic 3D scenes from a single image. While prior works are restricted to rigid body or simple elastic dynamics, WonderPlay features a hybrid generative simulator to synthesize a wide range of 3D dynamics. The hybrid generative simulator first uses a physics solver to simulate coarse 3D dynamics, which subsequently conditions a video generator to produce a video with finer, more realistic motion. The generated video is then used to update the simulated dynamic 3D scene, closing the loop between the physics solver and the video generator. This approach enables intuitive user control to be combined with the accurate dynamics of physics-based simulators and the expressivity of diffusion-based video generators. Experimental results demonstrate that WonderPlay enables users to interact with various scenes of diverse content, including cloth, sand, snow, liquid, smoke, elastic, and rigid bodies -- all using a single image input. Code will be made public. Project website: https://kyleleey.github.io/WonderPlay/
We propose UniPhy, a common latent-conditioned neural constitutive model that can encode the physical properties of diverse materials. At inference UniPhy allows `inverse simulation' i.e. inferring material properties by optimizing the scene-specific latent to match the available observations via differentiable simulation. In contrast to existing methods that treat such inference as system identification, UniPhy does not rely on user-specified material type information. Compared to prior neural constitutive modeling approaches which learn instance specific networks, the shared training across materials improves both, robustness and accuracy of the estimates. We train UniPhy using simulated trajectories across diverse geometries and materials -- elastic, plasticine, sand, and fluids (Newtonian & non-Newtonian). At inference, given an object with unknown material properties, UniPhy can infer the material properties via latent optimization to match the motion observations, and can then allow re-simulating the object under diverse scenarios. We compare UniPhy against prior inverse simulation methods, and show that the inference from UniPhy enables more accurate replay and re-simulation under novel conditions.
Feature selection is a critical step in data-driven applications, reducing input dimensionality to enhance learning accuracy, computational efficiency, and interpretability. Existing state-of-the-art methods often require post-selection retraining and extensive hyperparameter tuning, complicating their adoption. We introduce a novel, non-intrusive feature selection layer that, given a target feature count $k$, automatically identifies and selects the $k$ most informative features during neural network training. Our method is uniquely simple, requiring no alterations to the loss function, network architecture, or post-selection retraining. The layer is mathematically elegant and can be fully described by: \begin{align} \nonumber \tilde{x}_i = a_i x_i + (1-a_i)z_i \end{align} where $x_i$ is the input feature, $\tilde{x}_i$ the output, $z_i$ a Gaussian noise, and $a_i$ trainable gain such that $\sum_i{a_i^2}=k$. This formulation induces an automatic clustering effect, driving $k$ of the $a_i$ gains to $1$ (selecting informative features) and the rest to $0$ (discarding redundant ones) via weighted noise distortion and gain normalization. Despite its extreme simplicity, our method delivers state-of-the-art performance on standard benchmark datasets and a novel real-world dataset, outperforming or matching existing approaches without requiring hyperparameter search for $k$ or retraining. Theoretical analysis in the context of linear regression further validates its efficacy. Our work demonstrates that simplicity and performance are not mutually exclusive, offering a powerful yet straightforward tool for feature selection in machine learning.
Air pollution remains one of the most pressing environmental challenges of the modern era, significantly impacting human health, ecosystems, and climate. While traditional air quality monitoring systems provide critical data, their high costs and limited spatial coverage hinder effective real-time pollutant identification. Recent advancements in micro-sensor technology have improved data collection but still lack efficient methods for source identification. This paper explores the innovative application of machine learning (ML) models to classify pollutants in real-time using only data from optical micro-sensors. We propose a novel classification framework capable of distinguishing between four pollutant scenarios: Background Pollution, Ash, Sand, and Candle. Three Machine Learning (ML) approaches - XGBoost, Long Short-Term Memory networks, and Hidden Markov Chains - are evaluated for their effectiveness in sequence modeling and pollutant identification. Our results demonstrate the potential of leveraging micro-sensors and ML techniques to enhance air quality monitoring, offering actionable insights for urban planning and environmental protection.
We present a snapshot imaging technique for recovering 3D surrounding views of miniature scenes. Due to their intricacy, miniature scenes with objects sized in millimeters are difficult to reconstruct, yet miniatures are common in life and their 3D digitalization is desirable. We design a catadioptric imaging system with a single camera and eight pairs of planar mirrors for snapshot 3D reconstruction from a dollhouse perspective. We place paired mirrors on nested pyramid surfaces for capturing surrounding multi-view images in a single shot. Our mirror design is customizable based on the size of the scene for optimized view coverage. We use the 3D Gaussian Splatting (3DGS) representation for scene reconstruction and novel view synthesis. We overcome the challenge posed by our sparse view input by integrating visual hull-derived depth constraint. Our method demonstrates state-of-the-art performance on a variety of synthetic and real miniature scenes.
Linear representation hypothesis posits that high-level concepts are encoded as linear directions in the representation spaces of LLMs. Park et al. (2024) formalize this notion by unifying multiple interpretations of linear representation, such as 1-dimensional subspace representation and interventions, using a causal inner product. However, their framework relies on single-token counterfactual pairs and cannot handle ambiguous contrasting pairs, limiting its applicability to complex or context-dependent concepts. We introduce a new notion of binary concepts as unit vectors in a canonical representation space, and utilize LLMs' (neural) activation differences along with maximum likelihood estimation (MLE) to compute concept directions (i.e., steering vectors). Our method, Sum of Activation-base Normalized Difference (SAND), formalizes the use of activation differences modeled as samples from a von Mises-Fisher (vMF) distribution, providing a principled approach to derive concept directions. We extend the applicability of Park et al. (2024) by eliminating the dependency on unembedding representations and single-token pairs. Through experiments with LLaMA models across diverse concepts and benchmarks, we demonstrate that our lightweight approach offers greater flexibility, superior performance in activation engineering tasks like monitoring and manipulation.
In geotechnical engineering, constitutive models play a crucial role in describing soil behavior under varying loading conditions. Data-driven deep learning (DL) models offer a promising alternative for developing predictive constitutive models. When prediction is the primary focus, quantifying the predictive uncertainty of a trained DL model and communicating this uncertainty to end users is crucial for informed decision-making. This study proposes a recursive Bayesian neural network (rBNN) framework, which builds upon recursive feedforward neural networks (rFFNNs) by introducing generalized Bayesian inference for uncertainty quantification. A significant contribution of this work is the incorporation of a sliding window approach in rFFNNs, allowing the models to effectively capture temporal dependencies across load steps. The rBNN extends this framework by treating model parameters as random variables, with their posterior distributions inferred using generalized variational inference. The proposed framework is validated on two datasets: (i) a numerically simulated consolidated drained (CD) triaxial dataset employing a hardening soil model and (ii) an experimental dataset comprising 28 CD triaxial tests on Baskarp sand. Comparative analyses with LSTM, Bi-LSTM, and GRU models demonstrate that the deterministic rFFNN achieves superior predictive accuracy, attributed to its transparent structure and sliding window design. While the rBNN marginally trails in accuracy for the experimental case, it provides robust confidence intervals, addressing data sparsity and measurement noise in experimental conditions. The study underscores the trade-offs between deterministic and probabilistic approaches and the potential of rBNNs for uncertainty-aware constitutive modeling.
Predicting the lateral pile response is challenging due to the complexity of pile-soil interactions. Machine learning (ML) techniques have gained considerable attention for their effectiveness in non-linear analysis and prediction. This study develops an interpretable ML-based model for predicting p-y curves of monopile foundations. An XGBoost model was trained using a database compiled from existing research. The results demonstrate that the model achieves superior predictive accuracy. Shapley Additive Explanations (SHAP) was employed to enhance interpretability. The SHAP value distributions for each variable demonstrate strong alignment with established theoretical knowledge on factors affecting the lateral response of pile foundations.
This paper analyzes and compares patterns of U.S. domestic rail freight volumes during, and after the disruptions caused by the 2007-2009 Great Recession and the COVID-19 pandemic in 2020. Trends in rail and intermodal shipment data are examined in conjunction with economic indicators, focusing on the extent of drop and recovery of freight volumes of various commodities and intermodal shipments, and the lead/lag time with respect to economic drivers. While impacts from and the rebound from the Great Recessions were slow to develop, COVID-19 produced both profound disruptions in the freight market and rapid rebound, with important variations across commodity types. Energy-related commodities (i.e., coal, petroleum, and fracking sand), dropped during the pandemic while demand for other commodities (i.e., grain products and lumber, and intermodal freight). rebounded rapidly and in some cases grew. Overall rail freight experienced a rapid rebound following the precipitous drop in traffic in March and April 2020, achieving a near-full recovery in five months. As the recovery proceeded through 2020, intermodal flow, containers moving by rail for their longest overland trips, rebounded strongly, some exceeding 2019 levels. In contrast, rail flows during the Great Recession changed slowly with the onset and recovery, extending over multiple years. Pandemic response reflected the impacts of quick shutdowns and a rapid shift in consumer purchasing patterns. Results for the pandemic illustrate the resilience of U.S. rail freight industry and the multifaceted role it plays in the overall logistics system. Amid a challenging logistical environment, freight rail kept goods moving when other methods of transport were constrained.
Multilayer perceptron (MLP) networks are predominantly used to develop data-driven constitutive models for granular materials. They offer a compelling alternative to traditional physics-based constitutive models in predicting nonlinear responses of these materials, e.g., elasto-plasticity, under various loading conditions. To attain the necessary accuracy, MLPs often need to be sufficiently deep or wide, owing to the curse of dimensionality inherent in these problems. To overcome this limitation, we present an elasto-plasticity informed Chebyshev-based Kolmogorov-Arnold network (EPi-cKAN) in this study. This architecture leverages the benefits of KANs and augmented Chebyshev polynomials, as well as integrates physical principles within both the network structure and the loss function. The primary objective of EPi-cKAN is to provide an accurate and generalizable function approximation for non-linear stress-strain relationships, using fewer parameters compared to standard MLPs. To evaluate the efficiency, accuracy, and generalization capabilities of EPi-cKAN in modeling complex elasto-plastic behavior, we initially compare its performance with other cKAN-based models, which include purely data-driven parallel and serial architectures. Furthermore, to differentiate EPi-cKAN's distinct performance, we also compare it against purely data-driven and physics-informed MLP-based methods. Lastly, we test EPi-cKAN's ability to predict blind strain-controlled paths that extend beyond the training data distribution to gauge its generalization and predictive capabilities. Our findings indicate that, even with limited data and fewer parameters compared to other approaches, EPi-cKAN provides superior accuracy in predicting stress components and demonstrates better generalization when used to predict sand elasto-plastic behavior under blind triaxial axisymmetric strain-controlled loading paths.
Globally, vector-borne diseases are increasing in distribution and frequency, affecting humans, domestic animals and livestock, and wildlife. Science-based management and prevention of these diseases requires a sound understanding of the distribution and environmental requirements of the vectors and hosts involved in disease transmission. Integrated Species Distribution Models (ISDM) account for diverse data types through hierarchical modeling and represent a significant advancement in species distribution modeling that have not yet been leveraged in disease ecology. We used this approach, as implemented in the recently developed R package RISDM, to assess the distribution of the soft tick subspecies Ornithodoros turicata americanus. We created an ISDM for O. t. americanus, using systematically collected field data and historical records of this tick species in the southeastern US, to predict its distribution and assess potential correlations with environmental variables. Given the novelty of this method, we compared the results to a conventional Maxent SDM and validated the results through data partitioning using true skills statistics (TSS), sensitivity, and area under the ROC curve (AUC) metrics. We found that a combination of climatic variables describing seasonality and temperature extremes, along with the amount of sand in the soil, determined the predicted intensity of occurrence of this tick species. When projected in geographic space, this distribution model predicted 62% of Florida as suitable habitat for this tick species. The ISDM presented a higher TSS and AUC than the Maxent conventional model, while sensitivity was similar between both models. Our case example shows the utility of ISDMs in disease ecology studies and highlights the broad range of geographic suitability for this important disease vector.
Experiment 1. Rooting of quince hardwood cuttings: Rooting success was influenced by both the concentrations of IBA and the selection of rooting media. However, the control group (without IBA) notably enhanced rooting when compared to the various IBA concentrations. Cuttings in the control group (without IBA) and those planted in river sand exhibited notably high percentages of successful rooting, underscoring the importance of the selected planting medium. Experiment 2. Bench grafting of loquat: The success of grafting loquat cutting stocks varied based on grafting dates, types of cuttings, and concentrations of IBA. However, IBA at different concentrations did not have a significant impact. Notably, certain interactions such as grafting on February 20 with loquat stock cuttings, yielded higher percentages of successful graft bud sprouting. Experiment 3. Performance of grafting loquats onto different rootstocks: Grafting success was notably influenced by the selection of rootstock, with loquat rootstock demonstrating superior performance compared to quince. The highest significant levels of successful grafting were attained on February 20, underscoring the crucial role of grafting dates. Experiment 4. Impact of tree stock types on grafting success: Grafting success percentage was higher in loquat tree stock when compared to quince. The consistency of grafting success percentages across three dates underscores the significant influence of rootstock type. Experiment 5. Bench grafting of loquat cutting stocks: Graft bud sprout percentages exhibited variations, with loquat stock cuttings surpassing quince. Grafting success demonstrated a consistent increase from February 20 to March 30, underscoring the importance of selecting appropriate grafting dates.
Side-scan sonar (SSS) imagery presents unique challenges in the classification of man-made objects on the seafloor due to the complex and varied underwater environments. Historically, experts have manually interpreted SSS images, relying on conventional machine learning techniques with hand-crafted features. While Convolutional Neural Networks (CNNs) significantly advanced automated classification in this domain, they often fall short when dealing with diverse seafloor textures, such as rocky or ripple sand bottoms, where false positive rates may increase. Recently, Vision Transformers (ViTs) have shown potential in addressing these limitations by utilizing a self-attention mechanism to capture global information in image patches, offering more flexibility in processing spatial hierarchies. This paper rigorously compares the performance of ViT models alongside commonly used CNN architectures, such as ResNet and ConvNext, for binary classification tasks in SSS imagery. The dataset encompasses diverse geographical seafloor types and is balanced between the presence and absence of man-made objects. ViT-based models exhibit superior classification performance across f1-score, precision, recall, and accuracy metrics, although at the cost of greater computational resources. CNNs, with their inductive biases, demonstrate better computational efficiency, making them suitable for deployment in resource-constrained environments like underwater vehicles. Future research directions include exploring self-supervised learning for ViTs and multi-modal fusion to further enhance performance in challenging underwater environments.
This paper aims to verify the effectiveness of the software implementation of the proposed algorithm in creating computer-generated images of sand mixtures using a photograph of sand as an input and its effectiveness in converting digital pictures into sand-based images out of the mixtures it generated. The method of this paper is to visually compare the photographed image of the actual mixtures to its computer-generated counterpart to verify if the mixture generation produces results as expected and compare the computer-generated sand-based images with its source to verify image reproduction maintains same image content. The results of the mixture comparison shows that the actual and the computer-generated ones have similar overall shade and color. Still, the generated one has a rougher texture and higher contrast due to the method of inheriting visual features by pixel, not by individual sand particles. The comparison of the sand-based image and its source has demonstrated the software's ability to maintain the essence of its contents during conversion while replacing its texture with the visual properties of the generated sand mixture. The result have shown that the software implementation of the proposed algorithm can effectively use the images of sand to generate images of its mixtures and use those mixture images to convert a digital picture into a computer-generated sand-based image.
Elasmobranchs (shark sand rays) represent a critical component of marine ecosystems. Yet, they are experiencing global population declines and effective monitoring of populations is essential to their protection. Underwater stationary videos, such as those from Baited Remote Underwater Video Stations (BRUVS), are critical for understanding elasmobranch spatial ecology and abundance. However, processing these videos requires time-consuming manual analysis that can delay conservation. To address this challenge, we developed SharkTrack, a semi-automatic underwater video analysis software. SharkTrack uses Convolutional Neural Networks (CNN) and Multi-Object Tracking to automatically detect and track elasmobranchs and provides an annotation pipeline to manually classify elasmobranch species and compute species-specific MaxN (ssMaxN), the standard metric of relative abundance. When tested on BRUVS footage from locations unseen by the CNN model during training, SharkTrack computed ssMaxN with 89% accuracy over 207 hours of footage. The semi-automatic SharkTrack pipeline required two minutes of manual classification per hour of video, an estimated 95% reduction of manual analysis time compared to traditional methods. Furthermore, we demonstrate SharkTrack accuracy across diverse marine ecosystems and elasmobranch species, an advancement compared to previous models, which were limited to specific species or locations. SharkTrack applications extend beyond BRUVS, facilitating the analysis of any underwater stationary video. By making video analysis faster and more accessible, SharkTrack enables research and conservation organisations to monitor elasmobranch populations more efficiently, thereby improving conservation efforts. To further support these goals, we provide public access to the SharkTrack software.
We propose accelerating the simulation of Lagrangian dynamics, such as fluid flows, granular flows, and elastoplasticity, with neural-operator-based reduced-order modeling. While full-order approaches simulate the physics of every particle within the system, incurring high computation time for dense inputs, we propose to simulate the physics on sparse graphs constructed by sampling from the spatially discretized system. Our discretization-invariant reduced-order framework trains on any spatial discretizations and computes temporal dynamics on any sparse sampling of these discretizations through neural operators. Our proposed approach is termed Graph Informed Optimized Reduced-Order Modeling or \textit{GIOROM}. Through reduced order modeling, we ensure lower computation time by sparsifying the system by 6.6-32.0$\times$, while ensuring high-fidelity full-order inference via neural fields. We show that our model generalizes to a range of initial conditions, resolutions, and materials. The code and the demos are provided at \url{https://github.com/HrishikeshVish/GIOROM}
We present a novel, simple and widely applicable semi-supervised procedure for anomaly detection in industrial and IoT environments, SAnD (Simple Anomaly Detection). SAnD comprises 5 steps, each leveraging well-known statistical tools, namely; smoothing filters, variance inflation factors, the Mahalanobis distance, threshold selection algorithms and feature importance techniques. To our knowledge, SAnD is the first procedure that integrates these tools to identify anomalies and help decipher their putative causes. We show how each step contributes to tackling technical challenges that practitioners face when detecting anomalies in industrial contexts, where signals can be highly multicollinear, have unknown distributions, and intertwine short-lived noise with the long(er)-lived actual anomalies. The development of SAnD was motivated by a concrete case study from our industrial partner, which we use here to show its effectiveness. We also evaluate the performance of SAnD by comparing it with a selection of semi-supervised methods on public datasets from the literature on anomaly detection. We conclude that SAnD is effective, broadly applicable, and outperforms existing approaches in both anomaly detection and runtime.