Search

We consider the Steklov problem on differential $p$-forms defined by M. Karpukhin and present geometric eigenvalue bounds in the setting of warped product manifolds in various scenarios. In particular, we obtain Escobar type lower bounds for warped product manifolds with non-negative Ricci curvature and strictly convex boundary, and certain sharp bounds for hypersurfaces of revolution, among others. We compare and contrast the behaviour with known results in the case of functions (i.e., $0$-forms), highlighting the influence of the underlying topology on the spectrum for $p$-forms in general.

In recent years, natural language processing (NLP) models have demonstrated remarkable capabilities in various domains beyond traditional text generation. In this work, we introduce PeptideGPT, a protein language model tailored to generate protein sequences with distinct properties: hemolytic activity, solubility, and non-fouling characteristics. To facilitate a rigorous evaluation of these generated sequences, we established a comprehensive evaluation pipeline consisting of ideas from bioinformatics to retain valid proteins with ordered structures. First, we rank the generated sequences based on their perplexity scores, then we filter out those lying outside the permissible convex hull of proteins. Finally, we predict the structure using ESMFold and select the proteins with pLDDT values greater than 70 to ensure ordered structure. The properties of generated sequences are evaluated using task-specific classifiers - PeptideBERT and HAPPENN. We achieved an accuracy of 76.26% in hemolytic, 72.46% in non-hemolytic, 78.84% in non-fouling, and 68.06% in solubility protein generation. Our experimental results demonstrate the effectiveness of PeptideGPT in de novo protein design and underscore the potential of leveraging NLP-based approaches for paving the way for future innovations and breakthroughs in synthetic biology and bioinformatics. Codes, models, and data used in this study are freely available at: https://github.com/aayush-shah14/PeptideGPT.

We consider the magnetic Steklov eigenvalue problem on compact Riemannian manifolds with boundary for generic magnetic potentials and establish various results concerning the spectrum. We provide equivalent characterizations of magnetic Steklov operators which are unitarily equivalent to the classical Steklov operator and study bounds for the smallest eigenvalue. We prove a Cheeger-Jammes type lower bound for the first eigenvalue by introducing magnetic Cheeger constants. We also obtain an analogue of an upper bound for the first magnetic Neumann eigenvalue due to Colbois, El Soufi, Ilias and Savo. In addition, we compute the full spectrum in the case of the Euclidean $2$-ball and $4$-ball for a particular choice of magnetic potential given by Killing vector fields, and discuss the behavior. Finally, we establish a comparison result for the magnetic Steklov operator associated with the manifold and the square root of the magnetic Laplacian on the boundary, which generalizes the uniform geometric upper bounds for the difference of the corresponding eigenvalues in the non-magnetic case due to Colbois, Girouard and Hassannezhad.

Peptides are essential in biological processes and therapeutics. In this study, we introduce Multi-Peptide, an innovative approach that combines transformer-based language models with Graph Neural Networks (GNNs) to predict peptide properties. We combine PeptideBERT, a transformer model tailored for peptide property prediction, with a GNN encoder to capture both sequence-based and structural features. By employing Contrastive Language-Image Pre-training (CLIP), Multi-Peptide aligns embeddings from both modalities into a shared latent space, thereby enhancing the model's predictive accuracy. Evaluations on hemolysis and nonfouling datasets demonstrate Multi-Peptide's robustness, achieving state-of-the-art 86.185% accuracy in hemolysis prediction. This study highlights the potential of multimodal learning in bioinformatics, paving the way for accurate and reliable predictions in peptide-based research and applications.

Superlattices from twisted graphene mono- and bi-layer systems give rise to on-demand many-body states such as Mott insulators and unconventional superconductors. These phenomena are ascribed to a combination of flat bands and strong Coulomb interactions. However, a comprehensive understanding is lacking because the low-energy band structure strongly changes when the electron filling is varied. Here, we gain direct access to the filling-dependent low energy bands of twisted bilayer graphene (TBG) and twisted double bilayer graphene (TDBG) by applying micro-focused angle-resolved photoemission spectroscopy to in situ gated devices. Our findings for the two systems are in stark contrast: The doping dependent dispersion for TBG can be described in a simple model, combining a filling-dependent rigid band shift with a many-body related bandwidth change. In TDBG, on the other hand, we find a complex behaviour of the low-energy bands, combining non-monotonous bandwidth changes and tuneable gap openings. Our work establishes the extent of electric field tunability of the low energy electronic states in twisted graphene superlattices and can serve to underpin the theoretical understanding of the resulting phenomena., Comment: 26 pages, 4 main figures and 7 supplementary figures

We introduce an extension of the non-equilibrium dynamical mean field theory to incorporate the effects of static random disorder in the dynamics of a many-particle system by integrating out different disorder configurations resulting in an effective time-dependent density-density interaction. We use this method to study the non-equilibrium transient dynamics of a system described by the Fermi Anderson-Hubbard model following an interaction and disorder quench. The method recovers the solution of the disorder-free case for which the system exhibits qualitatively distinct dynamical behaviors in the weak-coupling (prethermalization) and strong-coupling regimes (collapse-and-revival oscillations). However, we find that weak random disorder promotes thermalization. In the weak coupling regime, the jump in the quasiparticle weight in the prethermal regime is suppressed by random disorder while in the strong-coupling regime, random disorder reduces the amplitude of the quasiparticle weight oscillations. These results highlight the importance of disorder in the dynamics of realistic many-particle systems., Comment: 6 pages, 3 figures

Retrieval-augmented generation (RAG) systems combine the strengths of language generation and information retrieval to power many real-world applications like chatbots. Use of RAG for understanding of videos is appealing but there are two critical limitations. One-time, upfront conversion of all content in large corpus of videos into text descriptions entails high processing times. Also, not all information in the rich video data is typically captured in the text descriptions. Since user queries are not known apriori, developing a system for video to text conversion and interactive querying of video data is challenging. To address these limitations, we propose an incremental RAG system called iRAG, which augments RAG with a novel incremental workflow to enable interactive querying of a large corpus of videos. Unlike traditional RAG, iRAG quickly indexes large repositories of videos, and in the incremental workflow, it uses the index to opportunistically extract more details from select portions of the videos to retrieve context relevant to an interactive user query. Such an incremental workflow avoids long video to text conversion times, and overcomes information loss issues due to conversion of video to text, by doing on-demand query-specific extraction of details in video data. This ensures high quality of responses to interactive user queries that are often not known apriori. To the best of our knowledge, iRAG is the first system to augment RAG with an incremental workflow to support efficient interactive querying of a large corpus of videos. Experimental results on real-world datasets demonstrate 23x to 25x faster video to text ingestion, while ensuring that latency and quality of responses to interactive user queries is comparable to responses from a traditional RAG where all video data is converted to text upfront before any user querying., Comment: Accepted in CIKM 2024

Automatic text summarization (TS) plays a pivotal role in condensing large volumes of information into concise, coherent summaries, facilitating efficient information retrieval and comprehension. This paper presents a novel framework for abstractive TS of single documents, which integrates three dominant aspects: structural, semantic, and neural-based approaches. The proposed framework merges machine learning and knowledge-based techniques to achieve a unified methodology. The framework consists of three main phases: pre-processing, machine learning, and post-processing. In the pre-processing phase, a knowledge-based Word Sense Disambiguation (WSD) technique is employed to generalize ambiguous words, enhancing content generalization. Semantic content generalization is then performed to address out-of-vocabulary (OOV) or rare words, ensuring comprehensive coverage of the input document. Subsequently, the generalized text is transformed into a continuous vector space using neural language processing techniques. A deep sequence-to-sequence (seq2seq) model with an attention mechanism is employed to predict a generalized summary based on the vector representation. In the post-processing phase, heuristic algorithms and text similarity metrics are utilized to refine the generated summary further. Concepts from the generalized summary are matched with specific entities, enhancing coherence and readability. Experimental evaluations conducted on prominent datasets, including Gigaword, Duc 2004, and CNN/DailyMail, demonstrate the effectiveness of the proposed framework. Results indicate significant improvements in handling rare and OOV words, outperforming existing state-of-the-art deep learning techniques. The proposed framework presents a comprehensive and unified approach towards abstractive TS, combining the strengths of structure, semantics, and neural-based methodologies.

The pursuit of novel alloys tailored to specific requirements poses significant challenges for researchers in the field. This underscores the importance of developing predictive techniques for essential physical properties of alloys based on their chemical composition and processing parameters. This study introduces AlloyBERT, a transformer encoder-based model designed to predict properties such as elastic modulus and yield strength of alloys using textual inputs. Leveraging the pre-trained RoBERTa encoder model as its foundation, AlloyBERT employs self-attention mechanisms to establish meaningful relationships between words, enabling it to interpret human-readable input and predict target alloy properties. By combining a tokenizer trained on our textual data and a RoBERTa encoder pre-trained and fine-tuned for this specific task, we achieved a mean squared error (MSE) of 0.00015 on the Multi Principal Elemental Alloys (MPEA) data set and 0.00611 on the Refractory Alloy Yield Strength (RAYS) dataset. This surpasses the performance of shallow models, which achieved a best-case MSE of 0.00025 and 0.0076 on the MPEA and RAYS datasets respectively. Our results highlight the potential of language models in material science and establish a foundational framework for text-based prediction of alloy properties that does not rely on complex underlying representations, calculations, or simulations., Comment: 20 pages, 3 figures

Intrinsically Disordered Proteins (IDPs) constitute a large and structure-less class of proteins with significant functions. The existence of IDPs challenges the conventional notion that the biological functions of proteins rely on their three-dimensional structures. Despite lacking well-defined spatial arrangements, they exhibit diverse biological functions, influencing cellular processes and shedding light on the disease mechanisms. However, it is expensive to run experiments or simulations to characterize this class of proteins. Consequently, we designed an ML model that relies solely on amino acid sequences. In this study, we introduce IDP-Bert model, a deep-learning architecture leveraging Transformers and Protein Language Models (PLMs) to map sequences directly to IDPs properties. Our experiments demonstrate accurate predictions of IDPs properties, including Radius of Gyration, end-to-end Decorrelation Time, and Heat Capacity., Comment: 22 pages, 5 figures

The out of time order correlator (OTOC) serves as a powerful tool for investigating quantum information spreading and chaos in complex systems. We present a method employing non-equilibrium dynamical mean-field theory (DMFT) and coherent potential approximation (CPA) combined with diagrammatic perturbation on the Schwinger-Keldysh contour to calculate the OTOC for correlated fermionic systems subjected to both random disorder and electrons interaction. Our key finding is that random disorder enhances the OTOC decay in the Hubbard model for the metallic phase in the weak coupling limit. However, the current limitation of our perturbative solver restricts the applicability to weak interaction regimes., Comment: 6 pages, 4 figures

Controlling the electronic structure of two-dimensional materials using the combination of twist angle and electrostatic doping is an effective means to induce emergent phenomena. In bilayer graphene with an interlayer twist angle near the magic angle, the electronic dispersion is strongly modified by a manifold of hybridizing moir\'e Dirac cones leading to flat band segments with strong electronic correlations. Numerous technical challenges arising from spatial inhomogeneity of interlayer interactions, twist angle and device functionality have so far limited momentum-resolved electronic structure measurements of these systems to static conditions. Here, we present a detailed characterization of the electronic structure exhibiting miniband dispersions for twisted bilayer graphene, near the magic angle, integrated in a functional device architecture using micro-focused angle-resolved photoemission spectroscopy. The optimum conditions for visualizing the miniband dispersion are determined by exploiting the spatial resolution and photon energy tunability of the light source and applied to extract a hybridization gap size of $(0.14 \pm 0.03)$~eV and flat band segments extending across a moir\'e mini Brillouin zone. \textit{In situ} electrostatic gating of the sample enables significant electron-doping, causing the conduction band states to shift below the Fermi energy. Our work emphasizes key challenges in probing the electronic structure of magic angle bilayer graphene devices and outlines conditions for exploring the doping-dependent evolution of the dispersion that underpins the ability to control many-body interactions in the material., Comment: 21 pages, 5 figures

This paper is based on developing different algorithms, which generate the task tree planning for the given goal node(recipe). The knowledge representation of the dishes is called FOON. It contains the different objects and their between them with respective to the motion node The graphical representation of FOON is made by noticing the change in the state of an object with respect to the human manipulators. We will explore how the FOON is created for different recipes by the robots. Task planning contains difficulties in exploring unknown problems, as its knowledge is limited to the FOON. To get the task tree planning for a given recipe, the robot will retrieve the information of different functional units from the knowledge retrieval process called FOON. Thus the generated subgraphs will allow the robot to cook the required dish. Thus the robot can able to cook the given recipe by following the sequence of instructions.

Ensuring high-quality video content for wireless users has become increasingly vital. Nevertheless, maintaining a consistent level of video quality faces challenges due to the fluctuating encoded bitrate, primarily caused by dynamic video content, especially in live streaming scenarios. Video compression is typically employed to eliminate unnecessary redundancies within and between video frames, thereby reducing the required bandwidth for video transmission. The encoded bitrate and the quality of the compressed video depend on encoder parameters, specifically, the quantization parameter (QP). Poor choices of encoder parameters can result in reduced bandwidth efficiency and high likelihood of non-conformance. Non-conformance refers to the violation of the peak signal-to-noise ratio (PSNR) constraint for an encoded video segment. To address these issues, a real-time deep learning-based H.264 controller is proposed. This controller dynamically estimates the optimal encoder parameters based on the content of a video chunk with minimal delay. The objective is to maintain video quality in terms of PSNR above a specified threshold while minimizing the average bitrate of the compressed video. Experimental results, conducted on both QCIF dataset and a diverse range of random videos from public datasets, validate the effectiveness of this approach. Notably, it achieves improvements of up to 2.5 times in average bandwidth usage compared to the state-of-the-art adaptive bitrate video streaming, with a negligible non-conformance probability below $10^{-2}$., Comment: arXiv admin note: text overlap with arXiv:2310.06857

Providing wireless users with high-quality video content has become increasingly important. However, ensuring consistent video quality poses challenges due to variable encoded bitrate caused by dynamic video content and fluctuating channel bitrate caused by wireless fading effects. Suboptimal selection of encoder parameters can lead to video quality loss due to underutilized bandwidth or the introduction of video artifacts due to packet loss. To address this, a real-time deep learning based H.264 controller is proposed. This controller leverages instantaneous channel quality data driven from the physical layer, along with the video chunk, to dynamically estimate the optimal encoder parameters with a negligible delay in real-time. The objective is to maintain an encoded video bitrate slightly below the available channel bitrate. Experimental results, conducted on both QCIF dataset and a diverse selection of random videos from public datasets, validate the effectiveness of the approach. Remarkably, improvements of 10-20 dB in PSNR with repect to the state-of-the-art adaptive bitrate video streaming is achieved, with an average packet drop rate as low as 0.002.

JPEG remains one of the most widespread lossy image coding methods. However, the non-differentiable nature of JPEG restricts the application in deep learning pipelines. Several differentiable approximations of JPEG have recently been proposed to address this issue. This paper conducts a comprehensive review of existing diff. JPEG approaches and identifies critical details that have been missed by previous methods. To this end, we propose a novel diff. JPEG approach, overcoming previous limitations. Our approach is differentiable w.r.t. the input image, the JPEG quality, the quantization tables, and the color conversion parameters. We evaluate the forward and backward performance of our diff. JPEG approach against existing methods. Additionally, extensive ablations are performed to evaluate crucial design choices. Our proposed diff. JPEG resembles the (non-diff.) reference implementation best, significantly surpassing the recent-best diff. approach by $3.47$dB (PSNR) on average. For strong compression rates, we can even improve PSNR by $9.51$dB. Strong adversarial attack results are yielded by our diff. JPEG, demonstrating the effective gradient approximation. Our code is available at https://github.com/necla-ml/Diff-JPEG., Comment: Accepted at WACV 2024. Project page: https://christophreich1996.github.io/differentiable_jpeg/ WACV paper: https://openaccess.thecvf.com/content/WACV2024/html/Reich_Differentiable_JPEG_The_Devil_Is_in_the_Details_WACV_2024_paper.html

Question-answering (QA) is a significant application of Large Language Models (LLMs), shaping chatbot capabilities across healthcare, education, and customer service. However, widespread LLM integration presents a challenge for small businesses due to the high expenses of LLM API usage. Costs rise rapidly when domain-specific data (context) is used alongside queries for accurate domain-specific LLM responses. One option is to summarize the context by using LLMs and reduce the context. However, this can also filter out useful information that is necessary to answer some domain-specific queries. In this paper, we shift from human-oriented summarizers to AI model-friendly summaries. Our approach, LeanContext, efficiently extracts $k$ key sentences from the context that are closely aligned with the query. The choice of $k$ is neither static nor random; we introduce a reinforcement learning technique that dynamically determines $k$ based on the query and context. The rest of the less important sentences are reduced using a free open source text reduction method. We evaluate LeanContext against several recent query-aware and query-unaware context reduction approaches on prominent datasets (arxiv papers and BBC news articles). Despite cost reductions of $37.29\%$ to $67.81\%$, LeanContext's ROUGE-1 score decreases only by $1.41\%$ to $2.65\%$ compared to a baseline that retains the entire context (no summarization). Additionally, if free pretrained LLM-based summarizers are used to reduce context (into human consumable summaries), LeanContext can further modify the reduced context to enhance the accuracy (ROUGE-1 score) by $13.22\%$ to $24.61\%$., Comment: The paper is under review

Efficient catalyst screening necessitates predictive models for adsorption energy, a key property of reactivity. However, prevailing methods, notably graph neural networks (GNNs), demand precise atomic coordinates for constructing graph representations, while integrating observable attributes remains challenging. This research introduces CatBERTa, an energy prediction Transformer model using textual inputs. Built on a pretrained Transformer encoder, CatBERTa processes human-interpretable text, incorporating target features. Attention score analysis reveals CatBERTa's focus on tokens related to adsorbates, bulk composition, and their interacting atoms. Moreover, interacting atoms emerge as effective descriptors for adsorption configurations, while factors such as bond length and atomic properties of these atoms offer limited predictive contributions. By predicting adsorption energy from the textual representation of initial structures, CatBERTa achieves a mean absolute error (MAE) of 0.75 eV-comparable to vanilla Graph Neural Networks (GNNs). Furthermore, the subtraction of the CatBERTa-predicted energies effectively cancels out their systematic errors by as much as 19.3% for chemically similar systems, surpassing the error reduction observed in GNNs. This outcome highlights its potential to enhance the accuracy of energy difference predictions. This research establishes a fundamental framework for text-based catalyst property prediction, without relying on graph representations, while also unveiling intricate feature-property relationships., Comment: 32 pages, 5 figures

Standardized lossy video coding is at the core of almost all real-world video processing pipelines. Rate control is used to enable standard codecs to adapt to different network bandwidth conditions or storage constraints. However, standard video codecs (e.g., H.264) and their rate control modules aim to minimize video distortion w.r.t. human quality assessment. We demonstrate empirically that standard-coded videos vastly deteriorate the performance of deep vision models. To overcome the deterioration of vision performance, this paper presents the first end-to-end learnable deep video codec control that considers both bandwidth constraints and downstream deep vision performance, while adhering to existing standardization. We demonstrate that our approach better preserves downstream deep vision performance than traditional standard video coding., Comment: Accepted at CVPR 2024 Workshop on AI for Streaming (AIS)

Recent advances in Language Models have enabled the protein modeling community with a powerful tool since protein sequences can be represented as text. Specifically, by taking advantage of Transformers, sequence-to-property prediction will be amenable without the need for explicit structural data. In this work, inspired by recent progress in Large Language Models (LLMs), we introduce PeptideBERT, a protein language model for predicting three key properties of peptides (hemolysis, solubility, and non-fouling). The PeptideBert utilizes the ProtBERT pretrained transformer model with 12 attention heads and 12 hidden layers. We then finetuned the pretrained model for the three downstream tasks. Our model has achieved state of the art (SOTA) for predicting Hemolysis, which is a task for determining peptide's potential to induce red blood cell lysis. Our PeptideBert non-fouling model also achieved remarkable accuracy in predicting peptide's capacity to resist non-specific interactions. This model, trained predominantly on shorter sequences, benefits from the dataset where negative examples are largely associated with insoluble peptides. Codes, models, and data used in this study are freely available at: https://github.com/ChakradharG/PeptideBERT, Comment: 24 pages

Deep learning based joint source-channel coding (JSCC) has demonstrated significant advancements in data reconstruction compared to separate source-channel coding (SSCC). This superiority arises from the suboptimality of SSCC when dealing with finite block-length data. Moreover, SSCC falls short in reconstructing data in a multi-user and/or multi-resolution fashion, as it only tries to satisfy the worst channel and/or the highest quality data. To overcome these limitations, we propose a novel deep learning multi-resolution JSCC framework inspired by the concept of multi-task learning (MTL). This proposed framework excels at encoding data for different resolutions through hierarchical layers and effectively decodes it by leveraging both current and past layers of encoded data. Moreover, this framework holds great potential for semantic communication, where the objective extends beyond data reconstruction to preserving specific semantic attributes throughout the communication process. These semantic features could be crucial elements such as class labels, essential for classification tasks, or other key attributes that require preservation. Within this framework, each level of encoded data can be carefully designed to retain specific data semantics. As a result, the precision of a semantic classifier can be progressively enhanced across successive layers, emphasizing the preservation of targeted semantics throughout the encoding and decoding stages. We conduct experiments on MNIST and CIFAR10 dataset. The experiment with both datasets illustrates that our proposed method is capable of surpassing the SSCC method in reconstructing data with different resolutions, enabling the extraction of semantic features with heightened confidence in successive layers. This capability is particularly advantageous for prioritizing and preserving more crucial semantic features within the datasets.

Non-Hermitian topological phases have gained immense attention due to their potential to unlock novel features beyond Hermitian bounds. PT-symmetric (Parity Time-reversal symmetric) non-Hermitian models have been studied extensively over the past decade. In recent years, the topological properties of general non-Hermitian models, regardless of the balance between gains and losses, have also attracted vast attention. Here we propose a non-Hermitian second-order topological (SOT) insulator that hosts gapless corner states on a two-dimensional quasi-crystalline lattice (QL). We first construct a non-Hermitian extension of the Bernevig-Hughes-Zhang (BHZ) model on a QL generated by the Amman-Beenker (AB) tiling. This model has real spectra and supports helical edge states. Corner states emerge by adding a proper Wilson mass term that gaps out the edge states. We propose two variations of the mass term that result in fascinating characteristics. In the first variation, we obtain a purely real spectra for the second-order topological phase. In the latter, we get a complex spectra with corner states localized at only two corners. Our findings pave a path to engineering exotic SOT phases where corner states can be localized at designated corners.

5G-LAN is an enterprise local area network (LAN) that leverages 5G technology for wireless connectivity instead of WiFi. 5G technology is unique: it uses network slicing to distinguish customers in the same traffic class using new QoS technologies in the RF domain. This unique ability is not supported by most enterprise LANs, which rely primarily on DiffServ-like technologies that distinguish among traffic classes rather than customers. We first show that this mismatch in QoS between the 5G network and the LAN affects the accuracy of insights from the LAN-resident analytics applications. We systematically analyze the root causes of the QoS mismatch and propose a first-of-a-kind 5G-LAN orchestrator (5GLoR). 5GLoR is a middleware that applications can use to preserve the QoS of their 5G data streams through the enterprise LAN. 5GLoR periodically analyzes the status of the queues, provides suitable DSCP identifiers to the application, and installs relevant switch re-write rules (to change DSCP identifiers between switches) to continuously preserve the QoS of the 5G data through the LAN. 5GLoR improves the RTP frame level delay and inter-frame delay by 212\% and 122\%, respectively, for the WebRTC application. Additionally, with 5GLoR, the accuracy of two example applications (face detection and recognition) improved by 33\%, while the latency was reduced by about 25\%. Our experiments show that the performance (accuracy and latency) of applications on a 5G-LAN performs well with the proposed 5GLoR compared to the same applications on MEC. This is significant because 5G-LAN offers an order of magnitude more computing, networking, and storage resources to the applications than the resource-constrained MEC, and mature enterprise technologies can be used to deploy, manage, and update IoT applications., Comment: 8 pages

IoT sensors, especially video cameras, are ubiquitously deployed around the world to perform a variety of computer vision tasks in several verticals including retail, healthcare, safety and security, transportation, manufacturing, etc. To amortize their high deployment effort and cost, it is desirable to perform multiple video analytics tasks, which we refer to as Analytical Units (AUs), off the video feed coming out of every camera. In this paper, we first show that in a multi-AU setting, changing the camera setting has disproportionate impact on different AUs performance. In particular, the optimal setting for one AU may severely degrade the performance for another AU, and further the impact on different AUs varies as the environmental condition changes. We then present Elixir, a system to enhance the video stream quality for multiple analytics on a video stream. Elixir leverages Multi-Objective Reinforcement Learning (MORL), where the RL agent caters to the objectives from different AUs and adjusts the camera setting to simultaneously enhance the performance of all AUs. To define the multiple objectives in MORL, we develop new AU-specific quality estimator values for each individual AU. We evaluate Elixir through real-world experiments on a testbed with three cameras deployed next to each other (overlooking a large enterprise parking lot) running Elixir and two baseline approaches, respectively. Elixir correctly detects 7.1% (22,068) and 5.0% (15,731) more cars, 94% (551) and 72% (478) more faces, and 670.4% (4975) and 158.6% (3507) more persons than the default-setting and time-sharing approaches, respectively. It also detects 115 license plates, far more than the time-sharing approach (7) and the default setting (0).

Cameras are increasingly being deployed in cities, enterprises and roads world-wide to enable many applications in public safety, intelligent transportation, retail, healthcare and manufacturing. Often, after initial deployment of the cameras, the environmental conditions and the scenes around these cameras change, and our experiments show that these changes can adversely impact the accuracy of insights from video analytics. This is because the camera parameter settings, though optimal at deployment time, are not the best settings for good-quality video capture as the environmental conditions and scenes around a camera change during operation. Capturing poor-quality video adversely affects the accuracy of analytics. To mitigate the loss in accuracy of insights, we propose a novel, reinforcement-learning based system APT that dynamically, and remotely (over 5G networks), tunes the camera parameters, to ensure a high-quality video capture, which mitigates any loss in accuracy of video analytics. As a result, such tuning restores the accuracy of insights when environmental conditions or scene content change. APT uses reinforcement learning, with no-reference perceptual quality estimation as the reward function. We conducted extensive real-world experiments, where we simultaneously deployed two cameras side-by-side overlooking an enterprise parking lot (one camera only has manufacturer-suggested default setting, while the other camera is dynamically tuned by APT during operation). Our experiments demonstrated that due to dynamic tuning by APT, the analytics insights are consistently better at all times of the day: the accuracy of object detection video analytics application was improved on average by ~ 42%. Since our reward function is independent of any analytics task, APT can be readily used for different video analytics tasks.

It is a common practice to think of a video as a sequence of images (frames), and re-use deep neural network models that are trained only on images for similar analytics tasks on videos. In this paper, we show that this leap of faith that deep learning models that work well on images will also work well on videos is actually flawed. We show that even when a video camera is viewing a scene that is not changing in any human-perceptible way, and we control for external factors like video compression and environment (lighting), the accuracy of video analytics application fluctuates noticeably. These fluctuations occur because successive frames produced by the video camera may look similar visually, but these frames are perceived quite differently by the video analytics applications. We observed that the root cause for these fluctuations is the dynamic camera parameter changes that a video camera automatically makes in order to capture and produce a visually pleasing video. The camera inadvertently acts as an unintentional adversary because these slight changes in the image pixel values in consecutive frames, as we show, have a noticeably adverse impact on the accuracy of insights from video analytics tasks that re-use image-trained deep learning models. To address this inadvertent adversarial effect from the camera, we explore the use of transfer learning techniques to improve learning in video analytics tasks through the transfer of knowledge from learning on image analytics tasks. In particular, we show that our newly trained Yolov5 model reduces fluctuation in object detection across frames, which leads to better tracking of objects(40% fewer mistakes in tracking). Our paper also provides new directions and techniques to mitigate the camera's adversarial effect on deep learning models used for video analytics applications.

Heterostructures composed of the intrinsic magnetic topological insulator MnBi$_2$Te$_4$ and its non-magnetic counterpart Bi$_2$Te$_3$ host distinct surface electronic band structures depending on the stacking order and exposed termination. Here, we probe the ultrafast dynamical response of MnBi$_2$Te$_4$ and MnBi$_4$Te$_7$ following near-infrared optical excitation using time- and angle-resolved photoemission spectroscopy, and disentangle surface from bulk dynamics based on density functional theory slab calculations of the surface-projected electronic structure. We gain access to the out-of-equilibrium charge carrier populations of both MnBi$_2$Te$_4$ and Bi$_2$Te$_3$ surface terminations of MnBi$_4$Te$_7$, revealing an instantaneous occupation of states associated with the Bi$_2$Te$_3$ surface layer followed by carrier extraction into the adjacent MnBi$_2$Te$_4$ layers with a laser fluence-tunable delay of up to 350 fs. The ensuing thermal relaxation processes are driven by phonon scattering with significantly slower relaxation times in the magnetic MnBi$_2$Te$_4$ septuple layers. The observed competition between interlayer charge transfer and intralayer phonon scattering demonstrates a method to control ultrafast charge transfer processes in MnBi$_2$Te$_4$-based van der Waals compounds., Comment: This document (21 pages, 4 figures) is the Accepted Manuscript version of a Published Work that appeared in final form in Nano Lett. 2023, 23, 2, 414-421, Copyright {\copyright} 2023 American Chemical Society after peer review. To access the final edited and published work see https://doi.org/10.1021/acs.nanolett.2c03075

The main objective of the present work is to assess higher-order Entropic Lattice Boltzmann Method (ELBM) for separated and transitional flows without the use of any explicit turbulence model. For this, we chose to simulate two cases of T106 Low-Pressure Turbine (LPT) cascade -- T106A and T106C -- representing incompressible and compressible flow regimes respectively. These results are obtained using our company's in-house higher-order ELBM transonic solver. We have carried out two sets of simulations for both the test cases. One with a clean inlet and the other with an inlet disturbance given by white Gaussian noise superimposed on the inlet velocity. For the clean inlet case, the flow remains laminar on the entire blade surface for both the test cases. It undergoes transition on the suction side for the inlet disturbance case. The pressure coefficient for the T106A and the isentropic Mach number on the blade surface for the T106C matches well with the experimental results. Also, the qualitative comparison of flow features in terms of the Q-criterion is in good agreement with the earlier computational results reported in literature., Comment: 8 pages, published in AIAA conference

Edge computing and 5G have made it possible to perform analytics closer to the source of data and achieve super-low latency response times, which is not possible with centralized cloud deployment. In this paper, we present a novel fever-screening system, which uses edge machine learning techniques and leverages private 5G to accurately identify and screen individuals with fever in real-time. Particularly, we present deep-learning based novel techniques for fusion and alignment of cross-spectral visual and thermal data streams at the edge. Our novel Cross-Spectral Generative Adversarial Network (CS-GAN) synthesizes visual images that have the key, representative object level features required to uniquely associate objects across visual and thermal spectrum. Two key features of CS-GAN are a novel, feature-preserving loss function that results in high-quality pairing of corresponding cross-spectral objects, and dual bottleneck residual layers with skip connections (a new, network enhancement) to not only accelerate real-time inference, but to also speed up convergence during model training at the edge. To the best of our knowledge, this is the first technique that leverages 5G networks and limited edge resources to enable real-time feature-level association of objects in visual and thermal streams (30 ms per full HD frame on an Intel Core i7-8650 4-core, 1.9GHz mobile processor). To the best of our knowledge, this is also the first system to achieve real-time operation, which has enabled fever screening of employees and guests in arenas, theme parks, airports and other critical facilities. By leveraging edge computing and 5G, our fever screening system is able to achieve 98.5% accuracy and is able to process about 5X more people when compared to a centralized cloud deployment.

With the growth of 5G, Internet of Things (IoT), edge computing and cloud computing technologies, the infrastructure (compute and network) available to emerging applications (AR/VR, autonomous driving, industry 4.0, etc.) has become quite complex. There are multiple tiers of computing (IoT devices, near edge, far edge, cloud, etc.) that are connected with different types of networking technologies (LAN, LTE, 5G, MAN, WAN, etc.). Deployment and management of applications in such an environment is quite challenging. In this paper, we propose ROMA, which performs resource orchestration for microservices-based 5G applications in a dynamic, heterogeneous, multi-tiered compute and network fabric. We assume that only application-level requirements are known, and the detailed requirements of the individual microservices in the application are not specified. As part of our solution, ROMA identifies and leverages the coupling relationship between compute and network usage for various microservices and solves an optimization problem in order to appropriately identify how each microservice should be deployed in the complex, multi-tiered compute and network fabric, so that the end-to-end application requirements are optimally met. We implemented two real-world 5G applications in video surveillance and intelligent transportation system (ITS) domains. Through extensive experiments, we show that ROMA is able to save up to 90%, 55% and 44% compute and up to 80%, 95% and 75% network bandwidth for the surveillance (watchlist) and transportation application (person and car detection), respectively. This improvement is achieved while honoring the application performance requirements, and it is over an alternative scheme that employs a static and overprovisioned resource allocation strategy by ignoring the resource coupling relationships., Comment: Accepted at 2022 IEEE/IFIP Network Operations and Management Symposium

Applications can tailor a network slice by specifying a variety of QoS attributes related to application-specific performance, function or operation. However, some QoS attributes like guaranteed bandwidth required by the application do vary over time. For example, network bandwidth needs of video streams from surveillance cameras can vary a lot depending on the environmental conditions and the content in the video streams. In this paper, we propose a novel, dynamic QoS attribute prediction technique that assists any application to make optimal resource reservation requests at all times. Standard forecasting using traditional cost functions like MAE, MSE, RMSE, MDA, etc. don't work well because they do not take into account the direction (whether the forecasting of resources is more or less than needed), magnitude (by how much the forecast deviates, and in which direction), or frequency (how many times the forecast deviates from actual needs, and in which direction). The direction, magnitude and frequency have a direct impact on the application's accuracy of insights, and the operational costs. We propose a new, parameterized cost function that takes into account all three of them, and guides the design of a new prediction technique. To the best of our knowledge, this is the first work that considers time-varying application requirements and dynamically adjusts slice QoS requests to 5G networks in order to ensure a balance between application's accuracy and operational costs. In a real-world deployment of a surveillance video analytics application over 17 cameras, we show that our technique outperforms other traditional forecasting methods, and it saves 34% of network bandwidth (over a ~24 hour period) when compared to a static, one-time reservation.

The exponential growth in smart sensors and rapid progress in 5G networks is creating a world awash with data streams. However, a key barrier to building performant multi-sensor, distributed stream processing applications is high programming complexity. We propose DataX, a novel platform that improves programmer productivity by enabling easy exchange, transformations, and fusion of data streams. DataX abstraction simplifies the application's specification and exposes parallelism and dependencies among the application functions (microservices). DataX runtime automatically sets up appropriate data communication mechanisms, enables effortless reuse of microservices and data streams across applications, and leverages serverless computing to transform, fuse, and auto-scale microservices. DataX makes it easy to write, deploy and reliably operate distributed applications at scale. Synthesizing these capabilities into a single platform is substantially more transformative than any available stream processing system.