Search

Vesicles are important surrogate structures made up of multiple phospholipids and cholesterol distributed in the form of a lipid bilayer. Tubular vesicles can undergo pearling i.e., formation of beads on the liquid thread akin to the Rayleigh-Plateau instability. Previous studies have inspected the effects of surface tension on the pearling instabilities of single-component vesicles. In this study, we perform a linear stability analysis on a multicomponent cylindrical vesicle. We solve the Stokes equations along with the Cahn-Hilliard equations to develop the linearized dynamic equations governing the vesicle shape and surface concentration fields. This helps us show that multicomponent vesicles can undergo pearling, buckling, and wrinkling even in the absence of surface tension, which is a significantly different result from studies on single-component vesicles. This behaviour arises due to the competition between the free energies of phase separation, line tension, and bending for this multi-phospholipid system. We determine the conditions under which axisymmetric and non-axisymmetric modes are dominant, and supplement our results with an energy analysis that shows the sources for these instabilities. We further show that these trends qualitatively match recent experiments.

Diffusion Models (DMs) have emerged as powerful generative models with unprecedented image generation capability. These models are widely used for data augmentation and creative applications. However, DMs reflect the biases present in the training datasets. This is especially concerning in the context of faces, where the DM prefers one demographic subgroup vs others (eg. female vs male). In this work, we present a method for debiasing DMs without relying on additional data or model retraining. Specifically, we propose Distribution Guidance, which enforces the generated images to follow the prescribed attribute distribution. To realize this, we build on the key insight that the latent features of denoising UNet hold rich demographic semantics, and the same can be leveraged to guide debiased generation. We train Attribute Distribution Predictor (ADP) - a small mlp that maps the latent features to the distribution of attributes. ADP is trained with pseudo labels generated from existing attribute classifiers. The proposed Distribution Guidance with ADP enables us to do fair generation. Our method reduces bias across single/multiple attributes and outperforms the baseline by a significant margin for unconditional and text-conditional diffusion models. Further, we present a downstream task of training a fair attribute classifier by rebalancing the training set with our generated data., Comment: CVPR 2024. Project Page : https://ab-34.github.io/balancing_act/

There have been significant advances in deep learning for music demixing in recent years. However, there has been little attention given to how these neural networks can be adapted for real-time low-latency applications, which could be helpful for hearing aids, remixing audio streams and live shows. In this paper, we investigate the various challenges involved in adapting current demixing models in the literature for this use case. Subsequently, inspired by the Hybrid Demucs architecture, we propose the Hybrid Spectrogram Time-domain Audio Separation Network HS-TasNet, which utilises the advantages of spectral and waveform domains. For a latency of 23 ms, the HS-TasNet obtains an overall signal-to-distortion ratio (SDR) of 4.65 on the MusDB test set, and increases to 5.55 with additional training data. These results demonstrate the potential of efficient demixing for real-time low-latency music applications., Comment: Accepted to ICASSP 2024

The application of Large Language Models (LLMs) in software engineering, particularly in static analysis tasks, represents a paradigm shift in the field. In this paper, we investigate the role that current LLMs can play in improving callgraph analysis and type inference for Python programs. Using the PyCG, HeaderGen, and TypeEvalPy micro-benchmarks, we evaluate 26 LLMs, including OpenAI's GPT series and open-source models such as LLaMA. Our study reveals that LLMs show promising results in type inference, demonstrating higher accuracy than traditional methods, yet they exhibit limitations in callgraph analysis. This contrast emphasizes the need for specialized fine-tuning of LLMs to better suit specific static analysis tasks. Our findings provide a foundation for further research towards integrating LLMs for static analysis tasks., Comment: To be published in: ICSE FORGE 2024 (AI Foundation Models and Software Engineering)

Experimental (design) optimization is a key driver in designing and discovering new products and processes. Bayesian Optimization (BO) is an effective tool for optimizing expensive and black-box experimental design processes. While Bayesian optimization is a principled data-driven approach to experimental optimization, it learns everything from scratch and could greatly benefit from the expertise of its human (domain) experts who often reason about systems at different abstraction levels using physical properties that are not necessarily directly measured (or measurable). In this paper, we propose a human-AI collaborative Bayesian framework to incorporate expert preferences about unmeasured abstract properties into the surrogate modeling to further boost the performance of BO. We provide an efficient strategy that can also handle any incorrect/misleading expert bias in preferential judgments. We discuss the convergence behavior of our proposed framework. Our experimental results involving synthetic functions and real-world datasets show the superiority of our method against the baselines., Comment: 19 Pages, 6 Figures

We propose and analyze a universal method to obtain fast charging of a quantum battery by a driven charger system using controlled, pure dephasing of the charger. While the battery displays coherent underdamped oscillations of energy for weak charger dephasing, the quantum Zeno freezing of the charger energy at high dephasing suppresses the rate of transfer of energy to the battery. Choosing an optimum dephasing rate between the regimes leads to a fast charging of the battery. We illustrate our results with the charger and battery modeled by either two-level systems or harmonic oscillators. Apart from the fast charging, the dephasing also renders the charging performance more robust to detuning between the charger, drive, and battery frequencies for the two-level systems case., Comment: 18 pages, 18 figures

In this paper, we outline potential ways for the further development of computational notebooks in Integrated Development Environments (IDEs). We discuss notebooks integration with IDEs, focusing on three main areas: facilitating experimentation, adding collaborative features, and improving code comprehension. We propose that better support of notebooks will not only benefit the notebooks, but also enhance IDEs by supporting new development processes native to notebooks. In conclusion, we suggest that adapting IDEs for more experimentation-oriented notebook processes will prepare them for the future of AI-powered programming.

Advances in artificial intelligence (AI) have enabled unprecedented capabilities, yet innovation teams struggle when envisioning AI concepts. Data science teams think of innovations users do not want, while domain experts think of innovations that cannot be built. A lack of effective ideation seems to be a breakdown point. How might multidisciplinary teams identify buildable and desirable use cases? This paper presents a first hand account of ideating AI concepts to improve critical care medicine. As a team of data scientists, clinicians, and HCI researchers, we conducted a series of design workshops to explore more effective approaches to AI concept ideation and problem formulation. We detail our process, the challenges we encountered, and practices and artifacts that proved effective. We discuss the research implications for improved collaboration and stakeholder engagement, and discuss the role HCI might play in reducing the high failure rate experienced in AI innovation., Comment: to appear at CHI 2024

Accurate triangulation of the domain plays a pivotal role in computing the numerical approximation of the differential operators. A good triangulation is the one which aids in reducing discretization errors. In a standard collocation technique, the smooth curved domain is typically triangulated with a mesh by taking points on the boundary to approximate them by polygons. However, such an approach often leads to geometrical errors which directly affect the accuracy of the numerical approximation. To restrict such geometrical errors, \textit{isoparametric}, \textit{subparametric}, and \textit{iso-geometric} methods were introduced which allow the approximation of the curved surfaces (or curved line segments). In this paper, we present an efficient finite element method to approximate the solution to the elliptic boundary value problem (BVP), which governs the response of an elastic solid containing a v-notch and inclusions. The algebraically nonlinear constitutive equation along with the balance of linear momentum reduces to second-order quasi-linear elliptic partial differential equation. Our approach allows us to represent the complex curved boundaries by smooth \textit{one-of-its-kind} point transformation. The main idea is to obtain higher-order shape functions which enable us to accurately compute the entries in the finite element matrices and vectors. A Picard-type linearization is utilized to handle the nonlinearities in the governing differential equation. The numerical results for the test cases show considerable improvement in the accuracy.

Reasoning from sequences of raw sensory data is a ubiquitous problem across fields ranging from medical devices to robotics. These problems often involve using long sequences of raw sensor data (e.g. magnetometers, piezoresistors) to predict sequences of desirable physical quantities (e.g. force, inertial measurements). While classical approaches are powerful for locally-linear prediction problems, they often fall short when using real-world sensors. These sensors are typically non-linear, are affected by extraneous variables (e.g. vibration), and exhibit data-dependent drift. For many problems, the prediction task is exacerbated by small labeled datasets since obtaining ground-truth labels requires expensive equipment. In this work, we present Hierarchical State-Space Models (HiSS), a conceptually simple, new technique for continuous sequential prediction. HiSS stacks structured state-space models on top of each other to create a temporal hierarchy. Across six real-world sensor datasets, from tactile-based state prediction to accelerometer-based inertial measurement, HiSS outperforms state-of-the-art sequence models such as causal Transformers, LSTMs, S4, and Mamba by at least 23% on MSE. Our experiments further indicate that HiSS demonstrates efficient scaling to smaller datasets and is compatible with existing data-filtering techniques. Code, datasets and videos can be found on https://hiss-csp.github.io.

Orthogonal Time Frequency Space (OTFS) is a framework for communications and active sensing that processes signals in the delay-Doppler (DD) domain. It is informed by 6G propagation environments, where Doppler spreads measured in kHz make it more and more difficult to estimate channels, and the standard model-dependent approach to wireless communication is starting to break down. We consider Zak-OTFS where inverse Zak transform converts information symbols mounted on DD domain pulses to the time domain for transmission. Zak-OTFS modulation is parameterized by a delay period $\tau_{p}$ and a Doppler period $\nu_{p}$, where the product $\tau_{p}\nu_{p}=1$. When the channel spread is less than the delay period, and the Doppler spread is less than the Doppler period, the Zak-OTFS input-output relation can be predicted from the response to a single pilot symbol. The highly reliable channel estimates concentrate around the pilot location, and we configure low-density parity-check (LDPC) codes that take advantage of this prior information about reliability. It is advantageous to allocate information symbols to more reliable bins in the DD domain. We report simulation results for a Veh-A channel model where it is not possible to resolve all the paths, showing that LDPC coding extends the range of Doppler spreads for which reliable model-free communication is possible. We show that LDPC coding reduces sensitivity to the choice of transmit filter, making bandwidth expansion less necessary. Finally, we compare BER performance of Zak-OTFS to that of a multicarrier approximation (MC-OTFS), showing LDPC coding amplifies the gains previously reported for uncoded transmission., Comment: 7 pages (double column), 6 figures, accepted at 2024 IEEE International Conference on Communications (ICC)

Background: Machine learning (ML) methods often fail with data that deviates from their training distribution. This is a significant concern for ML-enabled devices in clinical settings, where data drift may cause unexpected performance that jeopardizes patient safety. Method: We propose a ML-enabled Statistical Process Control (SPC) framework for out-of-distribution (OOD) detection and drift monitoring. SPC is advantageous as it visually and statistically highlights deviations from the expected distribution. To demonstrate the utility of the proposed framework for monitoring data drift in radiological images, we investigated different design choices, including methods for extracting feature representations, drift quantification, and SPC parameter selection. Results: We demonstrate the effectiveness of our framework for two tasks: 1) differentiating axial vs. non-axial computed tomography (CT) images and 2) separating chest x-ray (CXR) from other modalities. For both tasks, we achieved high accuracy in detecting OOD inputs, with 0.913 in CT and 0.995 in CXR, and sensitivity of 0.980 in CT and 0.984 in CXR. Our framework was also adept at monitoring data streams and identifying the time a drift occurred. In a simulation with 100 daily CXR cases, we detected a drift in OOD input percentage from 0-1% to 3-5% within two days, maintaining a low false-positive rate. Through additional experimental results, we demonstrate the framework's data-agnostic nature and independence from the underlying model's structure. Conclusion: We propose a framework for OOD detection and drift monitoring that is agnostic to data, modality, and model. The framework is customizable and can be adapted for specific applications.

Remote labs are a groundbreaking development in the education industry, providing students with access to laboratory education anytime, anywhere. However, most remote labs are costly and difficult to scale, especially in developing countries. With this as a motivation, this paper proposes a new remote labs (RLabs) solution that includes two use case experiments: Vanishing Rod and Focal Length. The hardware experiments are built at a low-cost by retrofitting Internet of Things (IoT) components. They are also made portable by designing miniaturised and modular setups. The software architecture designed as part of the solution seamlessly supports the scalability of the experiments, offering compatibility with a wide range of hardware devices and IoT platforms. Additionally, it can live-stream remote experiments without needing dedicated server space for the stream. The software architecture also includes an automation suite that periodically checks the status of the experiments using computer vision (CV). RLabs is qualitatively evaluated against seven non-functional attributes - affordability, portability, scalability, compatibility, maintainability, usability, and universality. Finally, user feedback was collected from a group of students, and the scores indicate a positive response to the students' learning and the platform's usability., Comment: 30 pages, 7 tables and 20 figures. Submitted to ACM Transactions on IoT

In this paper, we study the "dataset bias" problem from a statistical standpoint, and identify the main cause of the problem as the strong correlation between a class attribute u and a non-class attribute b in the input x, represented by p(u|b) differing significantly from p(u). Since p(u|b) appears as part of the sampling distributions in the standard maximum log-likelihood (MLL) objective, a model trained on a biased dataset via MLL inherently incorporates such correlation into its parameters, leading to poor generalization to unbiased test data. From this observation, we propose to mitigate dataset bias via either weighting the objective of each sample n by \frac{1}{p(u_{n}|b_{n})} or sampling that sample with a weight proportional to \frac{1}{p(u_{n}|b_{n})}. While both methods are statistically equivalent, the former proves more stable and effective in practice. Additionally, we establish a connection between our debiasing approach and causal reasoning, reinforcing our method's theoretical foundation. However, when the bias label is unavailable, computing p(u|b) exactly is difficult. To overcome this challenge, we propose to approximate \frac{1}{p(u|b)} using a biased classifier trained with "bias amplification" losses. Extensive experiments on various biased datasets demonstrate the superiority of our method over existing debiasing techniques in most settings, validating our theoretical analysis.

At the heart of the flourishing field of machine learning potentials are graph neural networks, where deep learning is interwoven with physics-informed machine learning (PIML) architectures. Various PIML models, upon training with density functional theory (DFT) material structure-property datasets, have achieved unprecedented prediction accuracy for a range of molecular and material properties. A critical component in the learned graph representation of crystal structures in PIMLs is how the various fragments of the structure's graph are embedded in a neural network. Several of the state-of-art PIML models apply spherical harmonic functions. Such functions are based on the assumption that DFT computes the Coulomb potential of atom-atom interactions. However, DFT does not directly compute such potentials, but integrates the electron-atom potentials. We introduce the direct integration of the external potential (DIEP) methods which more faithfully reflects that actual computational workflow in DFT. DIEP integrates the external (electron-atom) potential and uses these quantities to embed the structure graph into a deep learning model. We demonstrate the enhanced accuracy of the DIEP model in predicting the energies of pristine and defective materials. By training DIEP to predict the potential energy surface, we show the ability of the model in predicting the onset of fracture of pristine and defective carbon nanotubes., Comment: 14 pages, 4 figures

As Generative AI rises in adoption, its use has expanded to include domains such as hiring and recruiting. However, without examining the potential of bias, this may negatively impact marginalized populations, including people with disabilities. To address this important concern, we present a resume audit study, in which we ask ChatGPT (specifically, GPT-4) to rank a resume against the same resume enhanced with an additional leadership award, scholarship, panel presentation, and membership that are disability related. We find that GPT-4 exhibits prejudice towards these enhanced CVs. Further, we show that this prejudice can be quantifiably reduced by training a custom GPTs on principles of DEI and disability justice. Our study also includes a unique qualitative analysis of the types of direct and indirect ableism GPT-4 uses to justify its biased decisions and suggest directions for additional bias mitigation work. Additionally, since these justifications are presumably drawn from training data containing real-world biased statements made by humans, our analysis suggests additional avenues for understanding and addressing human bias.

We propose an approach for continuous prediction of turn-taking and backchanneling locations in spoken dialogue by fusing a neural acoustic model with a large language model (LLM). Experiments on the Switchboard human-human conversation dataset demonstrate that our approach consistently outperforms the baseline models with single modality. We also develop a novel multi-task instruction fine-tuning strategy to further benefit from LLM-encoded knowledge for understanding the tasks and conversational contexts, leading to additional improvements. Our approach demonstrates the potential of combined LLMs and acoustic models for a more natural and conversational interaction between humans and speech-enabled AI agents., Comment: To appear in IEEE ICASSP 2024

We report efficient terahertz (THz) generation in epitaxial $\alpha$-W/Co$_{60}$Fe$_{40}$ spintronic emitters. Two types of emitters have been investigated; epitaxial $\alpha$-W$(110)$/Co$_{60}$Fe$_{40}(110)$ and $\alpha$-W$(001)$/Co$_{60}$Fe$_{40}(001)$ deposited on single crystalline Al$_{2}$O$_{3}$($11\bar{2}0$) and MgO($001$) substrates, respectively. First principle calculations of the electronic band structure at the W$(001)$ surface reveal Dirac-type surface states, similar to that reported previously for the W$(110)$ surface. The generated THz radiation is about $10\%$ larger for $\alpha$-W$(110)$/Co$_{60}$Fe$_{40}(110)$ grown on single crystalline Al$_{2}$O$_{3}$($11\bar{2}0$), which is explained by the fact that the $\alpha$-W$(110)$/Co$_{60}$Fe$_{40}(110)$ interface for this emitter is more transparent to the spin current due to the presence of \AA ngstr\" om-scale interface intermixing at the W/CoFe interface. Our results also reveal that the generation of THz radiation is larger when pumping with the laser light from the substrate side, which is explained by a larger part of the laser light due to interference effects in the film stack being absorbed in the ferromagnetic Co$_{60}$Fe$_{40}$ layer in this measurement configuration.

Endpoint (EP) detection is a key component of far-field speech recognition systems that assist the user through voice commands. The endpoint detector has to trade-off between accuracy and latency, since waiting longer reduces the cases of users being cut-off early. We propose a novel two-pass solution for endpointing, where the utterance endpoint detected from a first pass endpointer is verified by a 2nd-pass model termed EP Arbitrator. Our method improves the trade-off between early cut-offs and latency over a baseline endpointer, as tested on datasets including voice-assistant transactional queries, conversational speech, and the public SLURP corpus. We demonstrate that our method shows improvements regardless of the first-pass EP model used., Comment: ASRU 2023

Epilepsy is one of the most common neurological diseases globally (around 50 million people worldwide). Fortunately, up to 70% of people with epilepsy could live seizure-free if properly diagnosed and treated, and a reliable technique to monitor the onset of seizures could improve the quality of life of patients who are constantly facing the fear of random seizure attacks. The scalp-based EEG test, despite being the gold standard for diagnosing epilepsy, is costly, necessitates hospitalization, demands skilled professionals for operation, and is discomforting for users. In this paper, we propose EarSD, a novel lightweight, unobtrusive, and socially acceptable ear-worn system to detect epileptic seizure onsets by measuring the physiological signals from behind the user's ears. EarSD includes an integrated custom-built sensing-computing-communication PCB to collect and amplify the signals of interest, remove the noises caused by motion artifacts and environmental impacts, and stream the data wirelessly to the computer/mobile phone nearby, where data are uploaded to the host computer for further processing. We conducted both in-lab and in-hospital experiments with epileptic seizure patients who were hospitalized for seizure studies.

The primary goal of this project is to develop privacy-preserving machine learning model training techniques for fNIRS data. This project will build a local model in a centralized setting with both differential privacy (DP) and certified robustness. It will also explore collaborative federated learning to train a shared model between multiple clients without sharing local fNIRS datasets. To prevent unintentional private information leakage of such clients' private datasets, we will also implement DP in the federated learning setting., Comment: 15 pages, 5 figures, 3 tables

In cross-domain retrieval, a model is required to identify images from the same semantic category across two visual domains. For instance, given a sketch of an object, a model needs to retrieve a real image of it from an online store's catalog. A standard approach for such a problem is learning a feature space of images where Euclidean distances reflect similarity. Even without human annotations, which may be expensive to acquire, prior methods function reasonably well using unlabeled images for training. Our problem constraint takes this further to scenarios where the two domains do not necessarily share any common categories in training data. This can occur when the two domains in question come from different versions of some biometric sensor recording identities of different people. We posit a simple solution, which is to generate synthetic data to fill in these missing category examples across domains. This, we do via category preserving translation of images from one visual domain to another. We compare approaches specifically trained for this translation for a pair of domains, as well as those that can use large-scale pre-trained text-to-image diffusion models via prompts, and find that the latter can generate better replacement synthetic data, leading to more accurate cross-domain retrieval models. Our best SynCDR model can outperform prior art by up to 15\%. Code for our work is available at https://github.com/samarth4149/SynCDR ., Comment: Pre-print

In light of the growing interest in type inference research for Python, both researchers and practitioners require a standardized process to assess the performance of various type inference techniques. This paper introduces TypeEvalPy, a comprehensive micro-benchmarking framework for evaluating type inference tools. TypeEvalPy contains 154 code snippets with 845 type annotations across 18 categories that target various Python features. The framework manages the execution of containerized tools, transforms inferred types into a standardized format, and produces meaningful metrics for assessment. Through our analysis, we compare the performance of six type inference tools, highlighting their strengths and limitations. Our findings provide a foundation for further research and optimization in the domain of Python type inference., Comment: To be published in ICSE 2024

In this work, we present the physics-informed neural network (PINN) model applied particularly to dynamic problems in solid mechanics. We focus on forward and inverse problems. Particularly, we show how a PINN model can be used efficiently for material identification in a dynamic setting. In this work, we assume linear continuum elasticity. We show results for two-dimensional (2D) plane strain problem and then we proceed to apply the same techniques for a three-dimensional (3D) problem. As for the training data we use the solution based on the finite element method. We rigorously show that PINN models are accurate, robust and computationally efficient, especially as a surrogate model for material identification problems. Also, we employ state-of-the-art techniques from the PINN literature which are an improvement to the vanilla implementation of PINN. Based on our results, we believe that the framework we have developed can be readily adapted to computational platforms for solving multiple dynamic problems in solid mechanics.

Training large foundation models using self-supervised objectives on unlabeled data, followed by fine-tuning on downstream tasks, has emerged as a standard procedure. Unfortunately, the efficacy of this approach is often constrained by both limited fine-tuning compute and scarcity in labeled downstream data. We introduce Multimodal Attention Merging (MAM), an attempt that facilitates direct knowledge transfer from attention matrices of models rooted in high resource modalities, text and images, to those in resource-constrained domains, speech and audio, employing a zero-shot paradigm. MAM reduces the relative Word Error Rate (WER) of an Automatic Speech Recognition (ASR) model by up to 6.70%, and relative classification error of an Audio Event Classification (AEC) model by 10.63%. In cases where some data/compute is available, we present Learnable-MAM, a data-driven approach to merging attention matrices, resulting in a further 2.90% relative reduction in WER for ASR and 18.42% relative reduction in AEC compared to fine-tuning., Comment: 5 pages, 1 figure, ICASSP 2024 Workshop on Self-supervision in Audio, Speech and Beyond

This document describes a two-day meeting held for the Principal Investigators (PIs) of NSF CyberTraining grants. The report covers invited talks, panels, and six breakout sessions. The meeting involved over 80 PIs and NSF program managers (PMs). The lessons recorded in detail in the report are a wealth of information that could help current and future PIs, as well as NSF PMs, understand the future directions suggested by the PI community. The meeting was held simultaneously with that of the PIs of the NSF Cyberinfrastructure for Sustained Scientific Innovation (CSSI) program. This co-location led to two joint sessions: one with NSF speakers and the other on broader impact. Further, the joint poster and refreshment sessions benefited from the interactions between CSSI and CyberTraining PIs., Comment: 38 pages, 3 main sections and 2 Appendix sections, 2 figures, 19 tables; updated version: author corrections

This research focuses on enabling Northeastern University's Husky, a multi-modal quadrupedal robot, to navigate narrow paths akin to various animals in nature. The Husky is equipped with thrusters to stabilize its body during dynamic maneuvers, addressing challenges inherent in aerial-legged systems. The approach involves modeling the robot as HROM (Husky Reduced Model) and creating an optimal control framework using linearized dynamics for narrow path walking. The thesis introduces a gait scheduling method to generate an open-loop walking gait and validates these gaits through a high-fidelity Simscape simulation. Experimental results of the open-loop walking are presented, accompanied by potential directions for advancing this robotic system., Comment: 60 pages, 27 figures

Identifying root causes of anomalies in causal processes is vital across disciplines. Once identified, one can isolate the root causes and implement necessary measures to restore the normal operation. Causal processes are often modelled as graphs with entities being nodes and their paths/interconnections as edge. Existing work only consider the contribution of nodes in the generative process, thus can not attribute the outlier score to the edges of the mechanism if the anomaly occurs in the connections. In this paper, we consider both individual edge and node of each mechanism when identifying the root causes. We introduce a noisy functional causal model to account for this purpose. Then, we employ Bayesian learning and inference methods to infer the noises of the nodes and edges. We then represent the functional form of a target outlier leaf as a function of the node and edge noises. Finally, we propose an efficient gradient-based attribution method to compute the anomaly attribution scores which scales linearly with the number of nodes and edges. Experiments on simulated datasets and two real-world scenario datasets show better anomaly attribution performance of the proposed method compared to the baselines. Our method scales to larger graphs with more nodes and edges., Comment: Accepted AAAI 2024

Inspired by Chukars wing-assisted incline running (WAIR), in this work, we employ a high-fidelity model of our Husky Carbon quadrupedal-legged robot to walk over steep slopes of up to 45 degrees. Chukars use the aerodynamic forces generated by their flapping wings to manipulate ground contact forces and traverse steep slopes and even overhangs. By exploiting the thrusters on Husky, we employed a collocation approach to rapidly resolving the joint and thruster actions. Our approach uses a polynomial approximation of the reduced-order dynamics of Husky, called HROM, to quickly and efficiently find optimal control actions that permit high-slope walking without violating friction cone conditions., Comment: arXiv admin note: substantial text overlap with arXiv:2306.00179

Recent advances in 3D fabrication have allowed handling the memory bottlenecks for modern data-intensive applications by bringing the computation closer to the memory, enabling Near Memory Processing (NMP). Memory Centric Networks (MCN) are advanced memory architectures that use NMP architectures, where multiple stacks of the 3D memory units are equipped with simple processing cores, allowing numerous threads to execute concurrently. The performance of the NMP is crucially dependent upon the efficient task offloading and task-to-NMP allocation. Our work presents a multi-armed bandit (MAB) based approach in formulating an efficient resource allocation strategy for MCN. Most existing literature concentrates only on one application domain and optimizing only one metric, i.e., either execution time or power. However, our solution is more generic and can be applied to diverse application domains. In our approach, we deploy Upper Confidence Bound (UCB) policy to collect rewards and eventually use it for regret optimization. We study the following metrics: instructions per cycle, execution times, NMP core cache misses, packet latencies, and power consumption. Our study covers various applications from PARSEC and SPLASH2 benchmarks suite. The evaluation shows that the system's performance improves by ~11% on average and an average reduction in total power consumption by ~12%.

In Near Memory Processing (NMP), processing elements(PEs) are placed near the 3D memory, reducing unnecessary data transfers between the CPU and the memory. However, as the CPUs and the PEs of the NMP use a shared memory space, maintaining coherency between them is a challenge. Most current literature relies on maintaining coherence for fine-grained or coarse-grained instruction granularities for the offloaded code blocks. We understand that for most NMP-offloaded instructions, the coherence conflict is low, and waiting for the coherence transaction hinders the performance. We construct an analytical model for an existing coherence strategy called CONDA, which is within 4% accuracy. This model indicates the key parameters responsible - the granularity of offloaded code, probability of conflicts, transaction times, and commit time. This paper identifies the prospective optimizations using the analytical model for CONDA. It proposes a new coherence scheme called MRCN: Monitored Rollback Coherence for NMP. MRCN addresses the coherence issue while eliminating unnecessary re-executions with limited hardware overhead. The MRCN is evaluated on synthetic as well as Rodinia benchmarks. The analytical results are within 4% accuracy of the simulation results. The MRCN shows improvement of upto 25% over CONDA strategy for the same benchmark under different execution conditions.

The ability to understand physical dynamics is critical for agents to act in the world. Here, we use Counterfactual World Modeling (CWM) to extract vision structures for dynamics understanding. CWM uses a temporally-factored masking policy for masked prediction of video data without annotations. This policy enables highly effective "counterfactual prompting" of the predictor, allowing a spectrum of visual structures to be extracted from a single pre-trained predictor without finetuning on annotated datasets. We demonstrate that these structures are useful for physical dynamics understanding, allowing CWM to achieve the state-of-the-art performance on the Physion benchmark., Comment: ECCV 2024. Project page at: https://neuroailab.github.io/cwm-physics/

The paper provides a survey of the development of machine-learning techniques for video analysis. The survey provides a summary of the most popular deep learning methods used for human activity recognition. We discuss how popular architectures perform on standard datasets and highlight the differences from real-life datasets dominated by multiple activities performed by multiple participants over long periods. For real-life datasets, we describe the use of low-parameter models (with 200X or 1,000X fewer parameters) that are trained to detect a single activity after the relevant objects have been successfully detected. Our survey then turns to a summary of machine learning methods that are specifically developed for working with a small number of labeled video samples. Our goal here is to describe modern techniques that are specifically designed so as to minimize the amount of ground truth that is needed for training and testing video analysis systems. We provide summaries of the development of self-supervised learning, semi-supervised learning, active learning, and zero-shot learning for applications in video analysis. For each method, we provide representative examples.

Recent data-privacy laws have sparked interest in machine unlearning, which involves removing the effect of specific training samples from a learnt model as if they were never present in the original training dataset. The challenge of machine unlearning is to discard information about the ``forget'' data in the learnt model without altering the knowledge about the remaining dataset and to do so more efficiently than the naive retraining approach. To achieve this, we adopt a projected-gradient based learning method, named as Projected-Gradient Unlearning (PGU), in which the model takes steps in the orthogonal direction to the gradient subspaces deemed unimportant for the retaining dataset, so as to its knowledge is preserved. By utilizing Stochastic Gradient Descent (SGD) to update the model weights, our method can efficiently scale to any model and dataset size. We provide empirically evidence to demonstrate that our unlearning method can produce models that behave similar to models retrained from scratch across various metrics even when the training dataset is no longer accessible. Our code is available at https://github.com/hnanhtuan/projected_gradient_unlearning., Comment: Accepted to WACV 2024

DP-SGD has emerged as a popular method to protect personally identifiable information in deep learning applications. Unfortunately, DP-SGD's per-sample gradient clipping and uniform noise addition during training can significantly degrade model utility. To enhance the model's utility, researchers proposed various adaptive DP-SGD methods. However, we examine and discover that these techniques result in greater privacy leakage or lower accuracy than the traditional DP-SGD method, or a lack of evaluation on a complex data set such as CIFAR100. To address these limitations, we propose an Auto DP-SGD. Our method automates clipping threshold estimation based on the DL model's gradient norm and scales the gradients of each training sample without losing gradient information. This helps to improve the algorithm's utility while using a less privacy budget. To further improve accuracy, we introduce automatic noise multiplier decay mechanisms to decrease the noise multiplier after every epoch. Finally, we develop closed-form mathematical expressions using tCDP accountant for automatic noise multiplier and automatic clipping threshold estimation. Through extensive experimentation, we demonstrate that Auto DP-SGD outperforms existing SOTA DP-SGD methods in privacy and accuracy on various benchmark datasets. We also show that privacy can be improved by lowering the scale factor and using learning rate schedulers without significantly reducing accuracy. Specifically, Auto DP-SGD, when used with a step noise multiplier, improves accuracy by 3.20, 1.57, 6.73, and 1.42 for the MNIST, CIFAR10, CIFAR100, and AG News Corpus datasets, respectively. Furthermore, it obtains a substantial reduction in the privacy budget of 94.9, 79.16, 67.36, and 53.37 for the corresponding data sets., Comment: 25 pages single column, 2 figures