Search

Recently, the number of off-the-shelf Large Language Models (LLMs) has exploded with many open-source options. This creates a diverse landscape regarding both serving options (e.g., inference on local hardware vs remote LLM APIs) and model heterogeneous expertise. However, it is hard for the user to efficiently optimize considering operational cost (pricing structures, expensive LLMs-as-a-service for large querying volumes), efficiency, or even per-case specific measures such as response accuracy, bias, or toxicity. Also, existing LLM routing solutions focus mainly on cost reduction, with response accuracy optimizations relying on non-generalizable supervised training, and ensemble approaches necessitating output computation for every considered LLM candidate. In this work, we tackle the challenge of selecting the optimal LLM from a model pool for specific queries with customizable objectives. We propose PickLLM, a lightweight framework that relies on Reinforcement Learning (RL) to route on-the-fly queries to available models. We introduce a weighted reward function that considers per-query cost, inference latency, and model response accuracy by a customizable scoring function. Regarding the learning algorithms, we explore two alternatives: PickLLM router acting as a learning automaton that utilizes gradient ascent to select a specific LLM, or utilizing stateless Q-learning to explore the set of LLMs and perform selection with a $\epsilon$-greedy approach. The algorithm converges to a single LLM for the remaining session queries. To evaluate, we utilize a pool of four LLMs and benchmark prompt-response datasets with different contexts. A separate scoring function is assessing response accuracy during the experiment. We demonstrate the speed of convergence for different learning rates and improvement in hard metrics such as cost per querying session and overall response latency., Comment: This work has been accepted at the first Workshop on Scalable and Efficient Artificial Intelligence Systems (SEAS) held in conjunction with the 39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025, in Philadelphia, Pennsylvania, USA

In image-assisted minimally invasive surgeries (MIS), understanding surgical scenes is vital for real-time feedback to surgeons, skill evaluation, and improving outcomes through collaborative human-robot procedures. Within this context, the challenge lies in accurately detecting, segmenting, and estimating the depth of surgical scenes depicted in high-resolution images, while simultaneously reconstructing the scene in 3D and providing segmentation of surgical instruments along with detection labels for each instrument. To address this challenge, a novel Multi-Task Learning (MTL) network is proposed for performing these tasks concurrently. A key aspect of this approach involves overcoming the optimization hurdles associated with handling multiple tasks concurrently by integrating a Adversarial Weight Update into the MTL framework, the proposed MTL model achieves 3D reconstruction through the integration of segmentation, depth estimation, and object detection, thereby enhancing the understanding of surgical scenes, which marks a significant advancement compared to existing studies that lack 3D capabilities. Comprehensive experiments on the EndoVis2018 benchmark dataset underscore the adeptness of the model in efficiently addressing all three tasks, demonstrating the efficacy of the proposed techniques., Comment: 1. Notation Update: Added * for equal contribution, ensuring proper attribution. 2. Subsection Fix: Removed the `subsection` tag for Section 3.1 (no 3.2 existed), maintaining content but fixing hierarchy. 3. Text Additions: Added lines in Section 5 and Subsection 4.2 for clarity, with references for better context

Professional networks provide invaluable entree to opportunity through referrals and introductions. A rich literature shows they also serve to entrench and even exacerbate a status quo of privilege and disadvantage. Hiring platforms, equipped with the ability to nudge link formation, provide a tantalizing opening for beneficial structural change. We anticipate that key to this prospect will be the ability to estimate the likelihood of edge formation in an evolving graph. Outcome-indistinguishable prediction algorithms ensure that the modeled world is indistinguishable from the real world by a family of statistical tests. Omnipredictors ensure that predictions can be post-processed to yield loss minimization competitive with respect to a benchmark class of predictors for many losses simultaneously, with appropriate post-processing. We begin by observing that, by combining a slightly modified form of the online K29 star algorithm of Vovk (2007) with basic facts from the theory of reproducing kernel Hilbert spaces, one can derive simple and efficient online algorithms satisfying outcome indistinguishability and omniprediction, with guarantees that improve upon, or are complementary to, those currently known. This is of independent interest. We apply these techniques to evolving graphs, obtaining online outcome-indistinguishable omnipredictors for rich -- possibly infinite -- sets of distinguishers that capture properties of pairs of nodes, and their neighborhoods. This yields, inter alia, multicalibrated predictions of edge formation with respect to pairs of demographic groups, and the ability to simultaneously optimize loss as measured by a variety of social welfare functions.

We initiate the study of differentially private learning in the proportional dimensionality regime, in which the number of data samples $n$ and problem dimension $d$ approach infinity at rates proportional to one another, meaning that $d / n \to \delta$ as $n \to \infty$ for an arbitrary, given constant $\delta \in (0, \infty)$. This setting is significantly more challenging than that of all prior theoretical work in high-dimensional differentially private learning, which, despite the name, has assumed that $\delta = 0$ or is sufficiently small for problems of sample complexity $O(d)$, a regime typically considered "low-dimensional" or "classical" by modern standards in high-dimensional statistics. We provide sharp theoretical estimates of the error of several well-studied differentially private algorithms for robust linear regression and logistic regression, including output perturbation, objective perturbation, and noisy stochastic gradient descent, in the proportional dimensionality regime. The $1 + o(1)$ factor precision of our error estimates enables a far more nuanced understanding of the price of privacy of these algorithms than that afforded by existing, coarser analyses, which are essentially vacuous in the regime we consider. We incorporate several probabilistic tools that have not previously been used to analyze differentially private learning algorithms, such as a modern Gaussian comparison inequality and recent universality laws with origins in statistical physics.

In optimal quantum control, control landscape phase transitions (CLPTs) indicate sharp changes occurring in the set of optimal protocols, as a physical model parameter is varied. Here, we demonstrate the existence of a new class of CLPTs, associated with changes in the topological properties of the optimal level set in a two-qubit state-preparation problem. In particular, the distance distribution of control protocols sampled through stochastic homotopic dynamics reveals discontinuous changes in the number of connected components in the optimal level set, as a function of the protocol duration. We demonstrate how topological CLPTs can be detected in modern-day experiments., Comment: 11 pages, 9 figures

The scale and diversity of demonstration data required for imitation learning is a significant challenge. We present EgoMimic, a full-stack framework which scales manipulation via human embodiment data, specifically egocentric human videos paired with 3D hand tracking. EgoMimic achieves this through: (1) a system to capture human embodiment data using the ergonomic Project Aria glasses, (2) a low-cost bimanual manipulator that minimizes the kinematic gap to human data, (3) cross-domain data alignment techniques, and (4) an imitation learning architecture that co-trains on human and robot data. Compared to prior works that only extract high-level intent from human videos, our approach treats human and robot data equally as embodied demonstration data and learns a unified policy from both data sources. EgoMimic achieves significant improvement on a diverse set of long-horizon, single-arm and bimanual manipulation tasks over state-of-the-art imitation learning methods and enables generalization to entirely new scenes. Finally, we show a favorable scaling trend for EgoMimic, where adding 1 hour of additional hand data is significantly more valuable than 1 hour of additional robot data. Videos and additional information can be found at https://egomimic.github.io/

Follow-up conversations with virtual assistants (VAs) enable a user to seamlessly interact with a VA without the need to repeatedly invoke it using a keyword (after the first query). Therefore, accurate Device-directed Speech Detection (DDSD) from the follow-up queries is critical for enabling naturalistic user experience. To this end, we explore the notion of Large Language Models (LLMs) and model the first query when making inference about the follow-ups (based on the ASR-decoded text), via prompting of a pretrained LLM, or by adapting a binary classifier on top of the LLM. In doing so, we also exploit the ASR uncertainty when designing the LLM prompts. We show on the real-world dataset of follow-up conversations that this approach yields large gains (20-40% reduction in false alarms at 10% fixed false rejects) due to the joint modeling of the previous speech context and ASR uncertainty, compared to when follow-ups are modeled alone.

Peptide therapeutics, including macrocycles, peptide inhibitors, and bioactive linear peptides, play a crucial role in therapeutic development due to their unique physicochemical properties. However, predicting these properties remains challenging. While structure-based models primarily focus on local interactions, language models are capable of capturing global therapeutic properties of both modified and linear peptides. Protein language models like ESM-2, though effective for natural peptides, cannot however encode chemical modifications. Conversely, pre-trained chemical language models excel in representing small molecule properties but are not optimized for peptides. To bridge this gap, we introduce PepDoRA, a unified peptide representation model. Leveraging Weight-Decomposed Low-Rank Adaptation (DoRA), PepDoRA efficiently fine-tunes the ChemBERTa-77M-MLM on a masked language model objective to generate optimized embeddings for downstream property prediction tasks involving both modified and unmodified peptides. By tuning on a diverse and experimentally valid set of 100,000 modified, bioactive, and binding peptides, we show that PepDoRA embeddings capture functional properties of input peptides, enabling the accurate prediction of membrane permeability, non-fouling and hemolysis propensity, and via contrastive learning, target protein-specific binding. Overall, by providing a unified representation for chemically and biologically diverse peptides, PepDoRA serves as a versatile tool for function and activity prediction, facilitating the development of peptide therapeutics across a broad spectrum of applications.

We study a federated version of multi-objective optimization (MOO), where a single model is trained to optimize multiple objective functions. MOO has been extensively studied in the centralized setting but is less explored in federated or distributed settings. We propose FedCMOO, a novel communication-efficient federated multi-objective optimization (FMOO) algorithm that improves the error convergence performance of the model compared to existing approaches. Unlike prior works, the communication cost of FedCMOO does not scale with the number of objectives, as each client sends a single aggregated gradient, obtained using randomized SVD (singular value decomposition), to the central server. We provide a convergence analysis of the proposed method for smooth non-convex objective functions under milder assumptions than in prior work. In addition, we introduce a variant of FedCMOO that allows users to specify a preference over the objectives in terms of a desired ratio of the final objective values. Through extensive experiments, we demonstrate the superiority of our proposed method over baseline approaches.

We study high-probability convergence in online learning, in the presence of heavy-tailed noise. To combat the heavy tails, a general framework of nonlinear SGD methods is considered, subsuming several popular nonlinearities like sign, quantization, component-wise and joint clipping. In our work the nonlinearity is treated in a black-box manner, allowing us to establish unified guarantees for a broad range of nonlinear methods. For symmetric noise and non-convex costs we establish convergence of gradient norm-squared, at a rate $\widetilde{\mathcal{O}}(t^{-1/4})$, while for the last iterate of strongly convex costs we establish convergence to the population optima, at a rate $\mathcal{O}(t^{-\zeta})$, where $\zeta \in (0,1)$ depends on noise and problem parameters. Further, if the noise is a (biased) mixture of symmetric and non-symmetric components, we show convergence to a neighbourhood of stationarity, whose size depends on the mixture coefficient, nonlinearity and noise. Compared to state-of-the-art, who only consider clipping and require unbiased noise with bounded $p$-th moments, $p \in (1,2]$, we provide guarantees for a broad class of nonlinearities, without any assumptions on noise moments. While the rate exponents in state-of-the-art depend on noise moments and vanish as $p \rightarrow 1$, our exponents are constant and strictly better whenever $p < 6/5$ for non-convex and $p < 8/7$ for strongly convex costs. Experiments validate our theory, demonstrating noise symmetry in real-life settings and showing that clipping is not always the optimal nonlinearity, further underlining the value of a general framework., Comment: 34 pages, 5 figures

The development of a "Large Airfoil Model (LAM)," a transformative approach for answering technical questions on airfoil aerodynamics, requires a vast dataset and a model to leverage it. To build this foundation, a novel probabilistic machine learning approach, A Deep Airfoil Prediction Tool (ADAPT), has been developed. ADAPT makes uncertainty-aware predictions of airfoil pressure coefficient ($C_p$) distributions by harnessing experimental data and incorporating measurement uncertainties. By employing deep kernel learning, performing Gaussian Process Regression in a ten-dimensional latent space learned by a neural network, ADAPT effectively handles unstructured experimental datasets. In tandem, Airfoil Surface Pressure Information Repository of Experiments (ASPIRE), the first large-scale, open-source repository of airfoil experimental data has been developed. ASPIRE integrates century-old historical data with modern reports, forming an unparalleled resource of real-world pressure measurements. This addresses a critical gap left by prior repositories, which relied primarily on numerical simulations. Demonstrative results for three airfoils show that ADAPT accurately predicts $C_p$ distributions and aerodynamic coefficients across varied flow conditions, achieving a mean absolute error in enclosed area ($\text{MAE}_\text{enclosed}$) of 0.029. ASPIRE and ADAPT lay the foundation for a future, more elaborate LAM that is coupled with a large language model, which will then enable users to perform design tasks based on design questions rather than explicit technical inputs.

Spatiotemporal data consisting of timestamps, GPS coordinates, and IDs occurs in many settings. Modeling approaches for this type of data must address challenges in terms of sensor noise, uneven sampling rates, and non-persistent IDs. In this work, we characterize and forecast human mobility at scale with dynamic generalized linear models (DGLMs). We represent mobility data as occupancy counts of spatial cells over time and use DGLMs to model the occupancy counts for each spatial cell in an area of interest. DGLMs are flexible to varying numbers of occupancy counts across spatial cells, are dynamic, and easily incorporate daily and weekly seasonality in the aggregate-level behavior. Our overall approach is robust to various types of noise and scales linearly in the number of spatial cells, time bins, and agents. Our results show that DGLMs provide accurate occupancy count forecasts over a variety of spatial resolutions and forecast horizons. We also present scaling results for spatiotemporal data consisting of hundreds of millions of observations. Our approach is flexible to support several downstream applications, including characterizing human mobility, forecasting occupancy counts, and anomaly detection for aggregate-level behaviors., Comment: 12 pages, 12 figures

In cross-device federated learning (FL) with millions of mobile clients, only a small subset of clients participate in training in every communication round, and Federated Averaging (FedAvg) is the most popular algorithm in practice. Existing analyses of FedAvg usually assume the participating clients are independently sampled in each round from a uniform distribution, which does not reflect real-world scenarios. This paper introduces a theoretical framework that models client participation in FL as a Markov chain to study optimization convergence when clients have non-uniform and correlated participation across rounds. We apply this framework to analyze a more general and practical pattern: every client must wait a minimum number of $R$ rounds (minimum separation) before re-participating. We theoretically prove and empirically observe that increasing minimum separation reduces the bias induced by intrinsic non-uniformity of client availability in cross-device FL systems. Furthermore, we develop an effective debiasing algorithm for FedAvg that provably converges to the unbiased optimal solution under arbitrary minimum separation and unknown client availability distribution.

This paper addresses the challenge of autonomous excavation of challenging terrains, in particular those that are prone to jamming and inter-particle adhesion when tackled by a standard penetrate-drag-scoop motion pattern. Inspired by human excavation strategies, our approach incorporates oscillatory rotation elements -- including swivel, twist, and dive motions -- to break up compacted, tangled grains and reduce jamming. We also present an adaptive impedance control method, the Reactive Attractor Impedance Controller (RAIC), that adapts a motion trajectory to unexpected forces during loading in a manner that tracks a trajectory closely when loads are low, but avoids excessive loads when significant resistance is met. Our method is evaluated on four terrains using a robotic arm, demonstrating improved excavation performance across multiple metrics, including volume scooped, protective stop rate, and trajectory completion percentage.

Gapped fracton phases constitute a new class of quantum states of matter which connects to topological orders but does not fit easily into existing paradigms. They host unconventional features such as sub-extensive and robust ground state degeneracies as well as sensitivity to lattice geometry. We investigate the anisotropic $\mathbb{Z}_N$ Laplacian model [1] which can describe a family of fracton phases defined on arbitrary graphs. Focusing on representative geometries where the 3D lattices are extensions of 2D square, triangular, honeycomb and Kagome lattices into the third dimension, we study their ground state degeneracies and mobility of excitations, and examine their entanglement renormalization group (ERG) flows. All models show bifurcating behaviors under ERG but have distinct ERG flows sensitive to both $N$ and lattice geometry. In particular, we show that the anisotropic $\mathbb{Z}_N$ Laplacian models defined on the extensions of triangular and honeycomb lattices are equivalent when $N$ is coprime to $3$. We also point out that, in contrast to previous expectations, the model defined on the extension of Kagome lattice is robust against local perturbations if and only if $N$ is coprime to $6$.

The growing demand for AI training data has transformed data annotation into a global industry, but traditional approaches relying on human annotators are often time-consuming, labor-intensive, and prone to inconsistent quality. We propose the Model-in-the-Loop (MILO) framework, which integrates AI/ML models into the annotation process. Our research introduces a collaborative paradigm that leverages the strengths of both professional human annotators and large language models (LLMs). By employing LLMs as pre-annotation and real-time assistants, and judges on annotator responses, MILO enables effective interaction patterns between human annotators and LLMs. Three empirical studies on multimodal data annotation demonstrate MILO's efficacy in reducing handling time, improving data quality, and enhancing annotator experiences. We also introduce quality rubrics for flexible evaluation and fine-grained feedback on open-ended annotations. The MILO framework has implications for accelerating AI/ML development, reducing reliance on human annotation alone, and promoting better alignment between human and machine values.

Experiments conducted by NASA measured the pressure drop due to gas-liquid flow through a packed-bed reactor under microgravity conditions. From these experiments, we develop correlations for the gas-liquid $f_{gl}$ interphase drag in a two-fluid model (TFM). We use an Ergun-type closure for liquid-solid drag. Then, under a 1D flow assumption, $f_{gl}$ is the only unknown in the TFM. Using a data-driven approach, we determine $f_{gl}$ and correlate it (via composite fits) with the liquid and gas Reynolds numbers, $Re_{l}$ and $Re_{g}$, respectively, and the Suratman number $Su_{l}$. To validate the proposed $f_{gl}(Re_{l},Re_{g},Su_{l})$ closure, we perform two-dimensional transient simulations at microgravity conditions using ANSYS Fluent and employing an Euler-Euler formulation. We find good agreement between the simulations based on the proposed $f_{gl}$ closure and the experimental data., Comment: 25 pages, 5 figures, preprint mode; v2: minor updates/enhancements

We propose and analyze a deformation of the 3+1d lattice $\mathbb Z_2$ gauge theory that preserves the non-invertible Wegner duality symmetry at the self-dual point. We identify a frustration-free point along this deformation where there are nine exactly degenerate ground states (on a periodic cubic lattice) even at finite volume. One of these ground states is a trivial product state and the rest are the topologically-ordered ground states of the 3+1d toric code. We also prove that the frustration-free point is gapped in the thermodynamic limit. Our model, therefore, realizes a gapped phase with spontaneously broken Wegner duality symmetry. Furthermore, by imposing the Gauss law constraints energetically, all the above features can be realized on a tensor product Hilbert space. Finally, we discuss a generalization of this deformation to the 3+1d lattice $\mathbb Z_N$ gauge theory and conjecture the possible phase diagram., Comment: 47+1 pages, 4 figures

Efficient management of infrastructure systems is crucial for economic stability, sustainability, and public safety. However, infrastructure sustainment is challenging due to the vast scale of systems, stochastic deterioration of components, partial observability, and resource constraints. Decision-making strategies that rely solely on human judgment often result in suboptimal decisions over large scales and long horizons. While data-driven approaches like reinforcement learning offer promising solutions, their application has been limited by the lack of suitable simulation environments. We present InfraLib, an open-source modular and extensible framework that enables modeling and analyzing infrastructure management problems with resource constraints as sequential decision-making problems. The framework implements hierarchical, stochastic deterioration models, supports realistic partial observability, and handles practical constraints including cyclical budgets and component unavailability. InfraLib provides standardized environments for benchmarking decision-making approaches, along with tools for expert data collection and policy evaluation. Through case studies on both synthetic benchmarks and real-world road networks, we demonstrate InfraLib's ability to model diverse infrastructure management scenarios while maintaining computational efficiency at scale., Comment: Updated preprint under active review

Distributed Denial of Service (DDoS) attacks are a major concern in network security, as they overwhelm systems with excessive traffic, compromise sensitive data, and disrupt network services. Accurately detecting these attacks is crucial to protecting network infrastructure. Traditional approaches, such as single Convolutional Neural Networks (CNNs) or conventional Machine Learning (ML) algorithms like Decision Trees (DTs) and Support Vector Machines (SVMs), struggle to extract the diverse features needed for precise classification, resulting in suboptimal performance. This research addresses this gap by introducing a novel approach for DDoS attack detection. The proposed method combines three distinct CNN architectures: SA-Enabled CNN with XGBoost, SA-Enabled CNN with LSTM, and SA-Enabled CNN with Random Forest. Each model extracts features at multiple scales, while self-attention mechanisms enhance feature integration and relevance. The weighted ensemble approach ensures that both prominent and subtle features contribute to the final classification, improving adaptability to evolving attack patterns and novel threats. The proposed method achieves a precision of 98.71%, an F1-score of 98.66%, a recall of 98.63%, and an accuracy of 98.69%, outperforming traditional methods and setting a new benchmark in DDoS attack detection. This innovative approach addresses critical limitations in current models and advances the state of the art in network security., Comment: 19 pages, 3 tables, 9 figures

Heat transfer between graphene and water is pivotal for various applications, including solarthermal vapor generation and the advanced manufacturing of graphene-based hierarchical structures in solution. In this study, we employ a deep-neural network potential derived from ab initio molecular dynamics to conduct extensive simulations of single-layer graphenewater systems with different levels of oxidation (carbon/oxygen ratio) of the graphene layer. Remarkably, our findings reveal a one-order-of-magnitude enhancement in heat transfer upon oxidizing graphene with hydroxyl or epoxide groups at the graphene surface, underscoring the significant tunability of heat transfer within this system. Given the same oxidation ratio, more dispersed locations of functional groups on graphene surface leads to faster heat dissipation to water., Comment: 21 pages, 8 figures

Control landscape phase transitions (CLPTs) occur as abrupt changes in the cost function landscape upon varying a control parameter, and can be revealed by non-analytic points in statistical order parameters. A prime example are quantum speed limits (QSL) which mark the onset of controllability as the protocol duration is increased. Here we lay the foundations of an analytical theory for CLPTs by developing Dyson, Magnus, and cumulant expansions for the cost function that capture the behavior of CLPTs with a controlled precision. Using linear and quadratic stability analysis, we reveal that CLPTs can be associated with different types of instabilities of the optimal protocol. This allows us to explicitly relate CLPTs to critical structural rearrangements in the extrema of the control landscape: utilizing path integral methods from statistical field theory, we trace back the critical scaling of the order parameter at the QSL to the topological and geometric properties of the set of optimal protocols, such as the number of connected components and its dimensionality. We verify our predictions by introducing a numerical sampling algorithm designed to explore this optimal set via a homotopic stochastic update rule. We apply this new toolbox explicitly to analyze CLPTs in the single- and two-qubit control problems whose landscapes are analytically tractable, and compare the landscapes for bang-bang and continuous protocols. Our work provides the first steps towards a systematic theory of CLPTs and paves the way for utilizing statistical field theory techniques for generic complex control landscapes., Comment: 38 pages, 24 figures

Under-canopy agricultural robots require robust navigation capabilities to enable full autonomy but struggle with tight row turning between crop rows due to degraded GPS reception, visual aliasing, occlusion, and complex vehicle dynamics. We propose an imitation learning approach using diffusion policies to learn row turning behaviors from demonstrations provided by human operators or privileged controllers. Simulation experiments in a corn field environment show potential in learning this task with only visual observations and velocity states. However, challenges remain in maintaining control within rows and handling varied initial conditions, highlighting areas for future improvement., Comment: Accepted as Extended Abstract to the IEEE ICRA@40 2024

Autonomous lander missions on extraterrestrial bodies need to sample granular materials while coping with domain shifts, even when sampling strategies are extensively tuned on Earth. To tackle this challenge, this paper studies the few-shot scooping problem and proposes a vision-based adaptive scooping strategy that uses the deep kernel Gaussian process method trained with a novel meta-training strategy to learn online from very limited experience on out-of-distribution target terrains. Our Deep Kernel Calibration with Maximal Deployment Gaps (kCMD) strategy explicitly trains a deep kernel model to adapt to large domain shifts by creating simulated maximal deployment gaps from an offline training dataset and training models to overcome these deployment gaps during training. Employed in a Bayesian Optimization sequential decision-making framework, the proposed method allows the robot to perform high-quality scooping actions on out-of-distribution terrains after a few attempts, significantly outperforming non-adaptive methods proposed in the excavation literature as well as other state-of-the-art meta-learning methods. The proposed method also demonstrates zero-shot transfer capability, successfully adapting to the NASA OWLAT platform, which serves as a state-of-the-art simulator for potential future planetary missions. These results demonstrate the potential of training deep models with simulated deployment gaps for more generalizable meta-learning in high-capacity models. Furthermore, they highlight the promise of our method in autonomous lander sampling missions by enabling landers to overcome the deployment gap between Earth and extraterrestrial bodies., Comment: arXiv admin note: substantial text overlap with arXiv:2303.02893

The foundational capabilities of humanoid robots should include robustly standing, walking, and mimicry of whole and partial-body motions. This work introduces the Masked Humanoid Controller (MHC), which supports all of these capabilities by tracking target trajectories over selected subsets of humanoid state variables while ensuring balance and robustness against disturbances. The MHC is trained in simulation using a carefully designed curriculum that imitates partially masked motions from a library of behaviors spanning standing, walking, optimized reference trajectories, re-targeted video clips, and human motion capture data. It also allows for combining joystick-based control with partial-body motion mimicry. We showcase simulation experiments validating the MHC's ability to execute a wide variety of behaviors from partially-specified target motions. Moreover, we demonstrate sim-to-real transfer on the real-world Digit V3 humanoid robot. To our knowledge, this is the first instance of a learned controller that can realize whole-body control of a real-world humanoid for such diverse multi-modal targets., Comment: Website: https://masked-humanoid.github.io/mhc/

We present foundation language models developed to power Apple Intelligence features, including a ~3 billion parameter model designed to run efficiently on devices and a large server-based language model designed for Private Cloud Compute. These models are designed to perform a wide range of tasks efficiently, accurately, and responsibly. This report describes the model architecture, the data used to train the model, the training process, how the models are optimized for inference, and the evaluation results. We highlight our focus on Responsible AI and how the principles are applied throughout the model development.

In this paper, a Hamiltonian lattice formulation for 2+1D compact Maxwell-Chern-Simons theory is derived. We analytically solve this theory and demonstrate that the mass gap in the continuum limit matches the well-known continuum formula. Our formulation preserves topological features such as the quantization of the Chern-Simons level, the degeneracy of energy eigenstates, the non-trivial properties of Wilson loops, and the mutual and self statistics of anyons. This work lays the groundwork for future Hamiltonian-based simulations of Maxwell-Chern-Simons theory on classical and quantum computers.

Efficient order fulfillment is vital in the agricultural industry, particularly due to the seasonal nature of seed supply chains. This paper addresses the challenge of optimizing seed orders fulfillment in a centralized warehouse where orders are processed in waves, taking into account the unpredictable arrival of seed stocks and strict order deadlines. We model the wave scheduling problem as a Markov decision process and propose an adaptive hybrid tree search algorithm that combines Monte Carlo tree search with domain-specific knowledge to efficiently navigate the complex, dynamic environment of seed distribution. By leveraging historical data and stochastic modeling, our method enables forecast-informed scheduling decisions that balance immediate requirements with long-term operational efficiency. The key idea is that we can augment Monte Carlo tree search algorithm with problem-specific side information that dynamically reduces the number of candidate actions at each decision step to handle the large state and action spaces that render traditional solution methods computationally intractable. Extensive simulations with realistic parameters-including a diverse range of products, a high volume of orders, and authentic seasonal durations-demonstrate that the proposed approach significantly outperforms existing industry standard methods.

De gustibus non est disputandum ("there is no accounting for others' tastes") is a common Latin maxim describing how many solutions in life are determined by people's personal preferences. Many household tasks, in particular, can only be considered fully successful when they account for personal preferences such as the visual aesthetic of the scene. For example, setting a table could be optimized by arranging utensils according to traditional rules of Western table setting decorum, without considering the color, shape, or material of each object, but this may not be a completely satisfying solution for a given person. Toward this end, we present DegustaBot, an algorithm for visual preference learning that solves household multi-object rearrangement tasks according to personal preference. To do this, we use internet-scale pre-trained vision-and-language foundation models (VLMs) with novel zero-shot visual prompting techniques. To evaluate our method, we collect a large dataset of naturalistic personal preferences in a simulated table-setting task, and conduct a user study in order to develop two novel metrics for determining success based on personal preference. This is a challenging problem and we find that 50% of our model's predictions are likely to be found acceptable by at least 20% of people., Comment: 19 pages, 10 figures

The evolution of networking paradigms has led to the emergence of Information-Centric Networking (ICN), Content-centric networking (CCNx), and Named Data Networking (NDN). These innovative architectures move away from traditional host-centric models to focus on content-oriented approaches. This paper offers a succinct understanding and in-depth exploration of these revolutionary networking frameworks. It explores into their foundational concepts, technical principles, and the latest developments. Also, the paper examines the current standards that govern these architectures and discusses their practical applications in the real world. By providing a comprehensive overview, this study aims to highlight how ICN, CCNx, and NDN are reshaping the landscape of digital communications, emphasizing efficiency, security, and scalability in data handling and distribution across networks.

Tensor networks provide a natural language for non-invertible symmetries in general Hamiltonian lattice models. We use ZX-diagrams, which are tensor network presentations of quantum circuits, to define a non-invertible operator implementing the Wegner duality in 3+1d lattice $\mathbb{Z}_2$ gauge theory. The non-invertible algebra, which mixes with lattice translations, can be efficiently computed using ZX-calculus. We further deform the $\mathbb{Z}_2$ gauge theory while preserving the duality and find a model with nine exactly degenerate ground states on a torus, consistent with the Lieb-Schultz-Mattis-type constraint imposed by the symmetry. Finally, we provide a ZX-diagram presentation of the non-invertible duality operators (including non-invertible parity/reflection symmetries) of generalized Ising models based on graphs, encompassing the 1+1d Ising model, the three-spin Ising model, the Ashkin-Teller model, and the 2+1d plaquette Ising model. The mixing (or lack thereof) with spatial symmetries is understood from a unifying perspective based on graph theory., Comment: 72+1 pages, 11 (numbered) figures, 2 tables

Increasing content consumption by users and the expectation of a better Internet experience requires Internet service providers (ISPs) to expand the capacity of the access network continually. The ISPs have been demanding the participation of the content providers (CPs) in sharing the cost of upgrading the infrastructure. From CPs' perspective, investing in the ISP infrastructure, termed as \emph{public investment}, seems rational as it will boost their profit. However, the CPs can alternatively invest in making content delivery more efficient, termed as \emph{private investment}, as it also boosts their profit. Thus, in this work, we investigate this trade-off between public and private investment of the CPs for a net-neutral ISP. Specifically, we consider centralized decision and non-cooperative forms of interaction between CPs and an ISP and determine the optimum public and private investments of the CPs for each model. In the non-cooperative interaction, we find that at most one CP contributes to the public infrastructure, whereas all invest in their private infrastructure.

This paper presents Neural Visibility Field (NVF), a novel uncertainty quantification method for Neural Radiance Fields (NeRF) applied to active mapping. Our key insight is that regions not visible in the training views lead to inherently unreliable color predictions by NeRF at this region, resulting in increased uncertainty in the synthesized views. To address this, we propose to use Bayesian Networks to composite position-based field uncertainty into ray-based uncertainty in camera observations. Consequently, NVF naturally assigns higher uncertainty to unobserved regions, aiding robots to select the most informative next viewpoints. Extensive evaluations show that NVF excels not only in uncertainty quantification but also in scene reconstruction for active mapping, outperforming existing methods., Comment: Accepted to CVPR 2024. More details can be found at https://sites.google.com/view/nvf-cvpr24/

Large language models have demonstrated impressive performance on commonsense tasks; however, these tasks are often posed as multiple-choice questions, allowing models to exploit systematic biases. Commonsense is also inherently probabilistic with multiple correct answers. The purpose of "boiling water" could be making tea and cooking, but it also could be killing germs. Existing tasks do not capture the probabilistic nature of common sense. To this end, we present commonsense frame completion (CFC), a new generative task that evaluates common sense via multiple open-ended generations. We also propose a method of probabilistic evaluation that strongly correlates with human judgments. Humans drastically outperform strong language model baselines on our dataset, indicating this approach is both a challenging and useful evaluation of machine common sense., Comment: ACL 2024 Camera Ready

Federated Learning (FL) enables edge devices or clients to collaboratively train machine learning (ML) models without sharing their private data. Much of the existing work in FL focuses on efficiently learning a model for a single task. In this paper, we study simultaneous training of multiple FL models using a common set of clients. The few existing simultaneous training methods employ synchronous aggregation of client updates, which can cause significant delays because large models and/or slow clients can bottleneck the aggregation. On the other hand, a naive asynchronous aggregation is adversely affected by stale client updates. We propose FedAST, a buffered asynchronous federated simultaneous training algorithm that overcomes bottlenecks from slow models and adaptively allocates client resources across heterogeneous tasks. We provide theoretical convergence guarantees for FedAST for smooth non-convex objective functions. Extensive experiments over multiple real-world datasets demonstrate that our proposed method outperforms existing simultaneous FL approaches, achieving up to 46.0% reduction in time to train multiple tasks to completion., Comment: Accepted to UAI 2024

For exponential last passage percolation on the plane we analyse the probability that the point-to-line geodesic exhibits an atypically large transversal fluctuation at the endpoint as well as the probability that the point-to-point geodesic exhibits an atypically large transversal fluctuation at the halfway point. In particular, we show that $p^*_n(t)$, the probability that the point-to-line geodesic from the origin to the line $x+y=2n$ ends at $(n-t(2n)^{2/3}, n+t(2n)^{2/3})$ satisfies that $n^{2/3}p^*_n(t)=\exp(-(\frac{4}{3}+o(1))t^{3})$ for $t$ large and $p_{n,\frac{1}{2}}(t)$, the probability that the geodesic from the origin to the point $(n,n)$ passes through the point $(\frac{1}{2}n-tn^{2/3}, \frac{1}{2} n+tn^{2/3})$, satisfies $n^{2/3}p_{n,\frac{1}{2}}(t)=\exp(-(\frac{8}{3}+o(1))t^3)$ for $t$ large. The latter result solves a special case of a conjecture from Liu (PTRF, 2022)., Comment: 19 pages, 4 figures