We introduce Probe Pruning (PP), a novel framework for online, dynamic, structured pruning of Large Language Models (LLMs) applied in a batch-wise manner. PP leverages the insight that not all samples and tokens contribute equally to the model's output, and probing a small portion of each batch effectively identifies crucial weights, enabling tailored dynamic pruning for different batches. It comprises three main stages: probing, history-informed pruning, and full inference. In the probing stage, PP selects a small yet crucial set of hidden states, based on residual importance, to run a few model layers ahead. During the history-informed pruning stage, PP strategically integrates the probing states with historical states. Subsequently, it structurally prunes weights based on the integrated states and the PP importance score, a metric developed specifically to assess the importance of each weight channel in maintaining performance. In the final stage, full inference is conducted on the remaining weights. A major advantage of PP is its compatibility with existing models, as it operates without requiring additional neural network modules or fine-tuning. Comprehensive evaluations of PP on LLaMA-2/3 and OPT models reveal that even minimal probing—using just 1.5% of FLOPs—can substantially enhance the efficiency of structured pruning of LLMs. For instance, when evaluated on LLaMA-2-7B with WikiText2, PP achieves a 2.56 times lower ratio of performance degradation per unit of latency reduction compared to the state-of-the-art method at a 40\% pruning ratio.

Ensuring that generative AI systems align with human values is essential but challenging, especially when considering multiple human values and their potential trade-offs. Since human values can be personalized and dynamically change over time, the desirable levels of value alignment vary across different ethnic groups, industry sectors, and user cohorts. Within existing frameworks, it is hard to define human values and align AI systems accordingly across different directions simultaneously, such as harmlessness, helpfulness, and positiveness. To address this, we develop a novel, first-principle approach called Multi-Human-Value Alignment Palette (MAP), which navigates the alignment across multiple human values in a structured and reliable way. MAP formulates the alignment problem as an optimization task with user-defined constraints, which define human value targets. It can be efficiently solved via a primal-dual approach, which determines whether a user-defined alignment target is achievable and how to achieve it. We conduct a detailed theoretical analysis of MAP by quantifying the trade-offs between values, the sensitivity to constraints, the fundamental connection between multi-value alignment and sequential alignment, and proving that linear weighted rewards are sufficient for multi-value alignment. Extensive experiments demonstrate MAP's ability to align multiple values in a principled manner while delivering strong empirical performance across various tasks.

In Federated Learning, clients train models on local data and send updates to a central server, which aggregates them into a global model using a fusion algorithm. This collaborative yet privacy-preserving training comes at a cost. FL developers face significant challenges in attributing global model predictions to specific clients. Localizing responsible clients is a crucial step towards (a) excluding clients primarily responsible for incorrect predictions and (b) encouraging clients who contributed highquality models to continue participating in the future. Existing ML debugging approaches are inherently inapplicable as they are designed for single-model, centralized training. We introduce TraceFL, a fine-grained neuron provenance capturing mechanism that identifies clients responsible for a global model's prediction by tracking the flow of information from individual clients to the global model. Since inference on different inputs activates a different set of neurons of the global model, TraceFL dynamically quantifies the significance of the global model's neurons in a given prediction, identifying the most crucial neurons in the global model. It then maps them to the corresponding neurons in every participating client to determine each client's contribution, ultimately localizing the responsible client. We evaluate TraceFL on six datasets, including two real-world medical imaging datasets and four neural networks, including advanced models such as GPT. TraceFL achieves 99% accuracy in localizing the responsible client in FL tasks spanning both image and text classification tasks. At a time when state-of-the-art ML debugging approaches are mostly domain-specific (e.g., image classification only), TraceFL is the first technique to enable highly accurate automated reasoning across a wide range of FL applications.

Federated Learning (FL) is an approach for privacypreserving Machine Learning (ML), enabling model training across multiple clients without centralized data collection. Existing incentive solutions for traditional Federated Learning (FL) focus on individual contributions to a single global objective, neglecting the nuances of clustered personalization with multiple cluster-level models and the non-monetary incentives such as personalized model appeal for clients. In this paper, we first propose to treat incentivization and personalization as interrelated challenges and solve them with an incentive mechanism that fosters personalized learning. Additionally, current methods depend on an aggregator for client clustering, which is limited by a lack of access to clients' confidential information due to privacy constraints, leading to inaccurate clustering. To overcome this, we propose direct client involvement, allowing clients to indicate their cluster membership preferences based on data distribution and incentive-driven feedback. Our approach enhances the personalized model appeal for self-aware clients with high-quality data leading to their active and consistent participation. Our evaluation demonstrates significant improvements in test accuracy ($8-45 \%$), personalized model appeal ($3-38 \%$), and participation rates ($\mathbf{31 - 100 \%}$) over existing FL models, including those addressing data heterogeneity and personalization.

Large Language Models (LLMs) like ChatGPT and Llama have revolutionized natural language processing and search engine dynamics. However, these models incur exceptionally high computational costs. For instance, GPT-3 consists of 175 billion parameters, where inference demands billions of floating-point operations. Caching is a natural solution to reduce LLM inference costs on repeated queries. However, existing caching methods are incapable of finding semantic similarities among LLM queries nor do they operate effectively on contextual queries, leading to unacceptable false hit-and-miss rates. This paper introduces MeanCache, a user-centric semantic cache for LLM-based services that identifies semantically similar queries to determine cache hit or miss. Using MeanCache, the response to a user's semantically similar query can be retrieved from a local cache rather than re-querying the LLM, thus reducing costs, service provider load, and environmental impact. MeanCache leverages Federated Learning (FL) to collaboratively train a query similarity model without violating user privacy. By placing a local cache in each user's device and using FL, MeanCache reduces the latency, costs, and enhances model performance, resulting in lower false-hit rates. MeanCache also encodes context chains for every cached query, offering a simple yet highly effective mechanism to discern contextual query responses from standalone queries. Our experiments benchmarked against the state-of-the-art caching method reveal that MeanCache attains an approximately 17 % higher F -score and a 20 % increase in precision during semantic cache hit-and-miss decisions while performing even better on contextual queries. It also reduces the storage requirement by 83 % and accelerates semantic cache hit-and-miss decisions by 11 %.

The integration of large language models (LLMs) into robotic task planning has unlocked better reasoning capabilities for complex, long-horizon workflows. However, ensuring safety in LLM-driven plans remains a critical challenge, as these models often prioritize task completion over risk mitigation. This paper introduces SAFER (Safety-Aware Framework for Execution in Robotics), a multi-LLM framework designed to embed safety awareness into robotic task planning. SAFER employs a Safety Agent that operates alongside the primary task planner, providing safety feedback. Additionally, we introduce LLM-as-a-Judge, a novel metric leveraging LLMs as evaluators to quantify safety violations within generated task plans. Our framework integrates safety feedback at multiple stages of execution, enabling real-time risk assessment, proactive error correction, and transparent safety evaluation. We also integrate a control framework using Control Barrier Functions (CBFs) to ensure safety guarantees within SAFER’s task planning. We evaluated SAFER against state-of-the-art LLM planners on complex long-horizon tasks involving heterogeneous robotic agents, demonstrating its effectiveness in reducing safety violations while maintaining task efficiency. We also verify the task planner and safety planner through actual hardware experiments involving multiple robots and a human.

Xinran Wang, Enmao Diao, Qi Le, Jie Ding, Ali Anwar. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Large language models can generate rare but catastrophic outputs, such as harmful conversations or insecure code. Existing Reinforcement Learning from Human Feedback (RLHF) typically maximizes average reward, leaving high-risk tail events insufficiently controlled. We introduce Quantile‑Guided Alignment (QA), a framework that allows users to specify desired improvements at any quantile—individually or across multiple reward dimensions—thus shifting the distribution of outputs with finer control toward safer, more desirable outcomes. The method extends standard RLHF via an augmented reward formulation that enforces quantile constraints. Experiments on conversation and code‐generation tasks show that quantile alignment significantly enhances quality at targeted tails while maintaining overall performance. The results position QA as a principled route to risk‑calibrated language models with tail‑focused alignment.

Personalized Federated Learning (pFL) holds immense promise for tailoring machine learning models to individual users while preserving data privacy. However, achieving optimal performance in pFL often requires a careful balancing act between memory overhead costs and model accuracy. This paper delves into the trade-offs inherent in pFL, offering valuable insights for selecting the right algorithms for diverse real-world scenarios. We empirically evaluate ten prominent pFL techniques across various datasets and data splits, uncovering significant differences in their performance. Our study reveals interesting insights into how pFL methods that utilize personalized (local) aggregation exhibit the fastest convergence due to their efficiency in communication and computation. Conversely, fine-tuning methods face limitations in handling data heterogeneity and potential adversarial attacks while multi-objective learning methods achieve higher accuracy at the cost of additional training and resource consumption. Our study emphasizes the critical role of communication efficiency in scaling pFL, demonstrating how it can significantly affect resource usage in real-world deployments.

Large language models (LLMs) have demonstrated remarkable capabilities, yet they occasionally exhibit sycophantic behavior, generating responses that align with or agree with a user’s stated opinions or preferences, even when those opinions are incorrect or biased. This sycophantic tendency can undermine the trustworthiness and reliability of LLMs. This work proposes a novel approach to mitigate sycophancy in LLMs by fine-tuning them on a carefully curated dataset comprising prompts paired with sycophantic and non-sycophantic responses1. Our method leverages Direct Preference Optimization (DPO), which optimizes LLMs to generate responses that align with the preferred (non-sycophantic) outputs without requiring explicit reward modeling. We develop a dataset of 1000 prompts with sycophantic and non-sycophantic responses to fine-tune LLMs. Our approach achieves an average reduction of 85% in persona-based tests and 84% in preference-driven tests, demonstrating significant mitigation of sycophantic behaviors. Our findings pave the way for more trustworthy and reliable language models that can provide objective and unbiased responses, aligning with human preferences while maintaining factual accuracy.

Federated Learning (FL) is a collaborative machine learning framework that allows multiple users to train models utilizing their local data in a distributed manner. However, considerable statistical heterogeneity in local data across devices often leads to suboptimal model performance compared with independently and identically distributed (IID) data scenarios. In this paper, we introduce DynamicFL, a new FL framework that investigates the trade-offs between global model performance and communication costs for two widely adopted FL methods: Federated Stochastic Gradient Descent (FedSGD) and Federated Averaging (FedAvg). Our approach allocates diverse communication resources to clients based on their data statistical heterogeneity, considering communication resource constraints, and attains substantial performance enhancements compared to uniform communication resource allocation. Notably, our method bridges the gap between FedSGD and FedAvg, providing a flexible framework leveraging communication heterogeneity to address statistical heterogeneity in FL. Through extensive experiments, we demonstrate that DynamicFL surpasses current state-of-the-art methods with up to a 10% increase in model accuracy, demonstrating its adaptability and effectiveness in tackling data statistical heterogeneity challenges.

Collaborations among various entities, such as companies, research labs, AI agents, and edge devices, have become increasingly crucial for achieving machine learning tasks that cannot be accomplished by a single entity alone. This is likely due to factors such as security constraints, privacy concerns, and limitations in computation resources. As a result, Collaborative Learning has been gaining momentum. However, a significant challenge in practical applications of Collaborative Learning is how to effectively incentivize multiple entities to collaborate before any collaboration occurs. In this study, we propose ICL, an architectural framework for Incentivized Collaborative Learning, and provide insights into the critical issue of when and why incentives can improve collaboration performance. We showcase the concepts of ICL to specific use cases in federated learning, assisted learning, and multi-armed bandit, corroborating with both theoretical and experimental results.

Federated Learning (FL) has emerged as a powerful approach that enables collaborative distributed model training without the need for data sharing. However, FL grapples with inherent heterogeneity challenges leading to issues such as stragglers, dropouts, and performance variations. Selection of clients to run an FL instance is crucial, but existing strategies introduce biases and participation issues and do not consider resource efficiency. Communication and training acceleration solutions proposed to increase client participation also fall short due to the dynamic nature of system resources. We address these challenges in this paper by designing FLOAT, a novel framework designed to boost FL client resource awareness. FLOAT optimizes resource utilization dynamically for meeting training deadlines, and mitigates stragglers and dropouts through various optimization techniques; leading to enhanced model convergence and improved performance. FLOAT leverages multi-objective Reinforcement Learning with Human Feedback (RLHF) to automate the selection of the optimization techniques and their configurations, tailoring them to individual client resource conditions. Moreover, FLOAT seamlessly integrates into existing FL systems, maintaining non-intrusiveness and versatility for both asynchronous and synchronous FL settings. As per our evaluations, FLOAT increases accuracy by up to 53%, reduces client dropouts by up to 78×, and improves communication, computation, and memory utilization by up to 81×, 44×, and 20× respectively.

As the number of pre-trained machine learning (ML) models is growing exponentially, data reduction tools are not catching up. Existing data reduction techniques are not specifically designed for pre-trained model (PTM) dataset files. This is largely due to a lack of understanding of the patterns and characteristics of these datasets, especially those relevant to data reduction and compressibility. This paper presents the first, exhaustive analysis to date of PTM datasets on storage compressibility. Our analysis spans different types of data reduction and compression techniques, from hash-based data deduplication, data similarity detection, to dictionary-coding compression. Our analysis explores these techniques at three data granularity levels, from model layers, model chunks, to model parameters. We draw new observations that indicate that modern data reduction tools are not effective when handling PTM datasets. There is a pressing need for new compression methods that take into account PTMs' data characteristics for effective storage reduction. Motivated by our findings, we design Elf, a simple yet effective, error-bounded, lossy floating-point compression method. Elf transforms floating-point parameters in such a way that the common exponent field of the transformed parameters can be completely eliminated to save storage space. We develop Elves, a compression framework that integrates Elf along with several other data reduction methods. Elves uses the most effective method to compress PTMs that exhibit different patterns. Evaluation shows that Elves achieves an overall compression ratio of 1.52×, which is 1.31×, 1.32× and 1.29× higher than a general-purpose compressor (zstd), an error-bounded lossy compressor (SZ3), and the uniform model quantization, respectively, with negligible model accuracy loss.

The wide adoption of Docker containers for supporting agile and elastic enterprise applications has led to a broad proliferation of container images. The associated storage performance and capacity requirements place a high pressure on the infrastructure of container registries that store and distribute images and container storage systems on the Docker client side that manage image layers and store ephemeral data generated at container runtime. The storage demand is worsened by the large amount of duplicate data in images. Moreover, container storage systems that use Copy-on-Write (CoW) file systems as storage drivers exacerbate the redundancy. Exploiting the high file redundancy in real-world images is a promising approach to drastically reduce the growing storage requirements of container registries and improve the space efficiency of container storage systems. However, existing deduplication techniques significantly degrade the performance of both registries and container storage systems because of data reconstruction overhead as well as the deduplication cost. We propose DupHunter, an end-to-end deduplication scheme that deduplicates layers for both Docker registries and container storage systems while maintaining a high image distribution speed and container I/O performance. DupHunter is divided into three tiers: registry tier, middle tier, and client tier. Specifically, we first build a high-performance deduplication engine at the registry tier that not only natively deduplicates layers for space savings but also reduces layer restore overhead. Then, we use deduplication offloading at the middle tier to eliminate the redundant files from the client tier and avoid bringing deduplication overhead to the clients. To further reduce the data duplicates caused by CoWs and improve the container I/O performance, we utilize a container-aware storage system at the client tier that reserves space for each container and arranges the placement of files and their modifications on the disk to preserve locality. Under real workloads, DupHunter reduces storage space by up to 6.9× and reduces the GET layer latency up to 2.8× compared to the state-of-the-art. Moreover, DupHunter can improve the container I/O performance by up to 93% for reads and 64% for writes.

Federated Learning (FL) is a machine learning approach that addresses privacy and data transfer costs by computing data at the source. It’s particularly popular for Edge and IoT applications where the aggregator server of FL is in resource-capped edge data centers for reducing communication costs. Existing cloud-based aggregator solutions are resource-inefficient and expensive at the Edge, leading to low scalability and high latency. To address these challenges, this study compares prior and new aggregation methodologies under the changing demands of IoT and Edge applications. This work is the first to propose an adaptive FL aggregator at the Edge, enabling users to manage the cost and efficiency trade-off. An extensive comparative analysis demonstrates that the design improves scalability by up to 4$\times$, time efficiency by 8$\times$, and reduces costs by more than 2$\times$ compared to extant cloud-based static methodologies.

Federated Learning (FL) has emerged as a new paradigm for training machine learning models distributively without sacrificing data security and privacy. Learning models on edge devices such as mobile phones is one of the most common use cases for FL. However, Non-identical independent distributed (non-IID) data in edge devices easily leads to training failures. Especially, over-parameterized machine learning models can easily be over-fitted on such data, hence, resulting in inefficient federated learning and poor model performance. To overcome the over-fitting issue, we proposed an adaptive dynamic pruning approach for FL, which can dynamically slim the model by dropping out unimportant parameters, hence, preventing over-fittings. Since the machine learning model's parameters react differently for different training samples, adaptive dynamic pruning will evaluate the salience of the model's parameter according to the input training sample, and only retain the salient parameter's gradients when doing back-propagation. We performed comprehensive experiments to evaluate our approach. The results show that our approach by removing the redundant parameters in neural networks can significantly reduce the over-fitting issue and greatly improves the training efficiency. In particular, when training the ResNet-32 on CIFAR-10, our approach reduces the communication cost by 57%. We further demonstrate the inference acceleration capability of the proposed algorithm. Our approach reduces up to 50% FLOPs inference of DNNs on edge devices while maintaining the model's quality.

Cross-device Federated Learning (FL) enables training machine learning (ML) models on private data that is heterogeneously distributed over many IoT end devices without violating privacy requirements. Clients typically vary significantly in data quality, hardware resources and stability, which results in challenges such as increased training times, higher resource costs, sub-par model performance and biased training. Existing works tend to address each of these challenges in isolation, but overlook how they might impact each other holistically. We perform a first of its kind characterization study that empirically demonstrates how these properties interact with each other to impact important performance metrics such as model error, fairness, resource cost and training time. We then propose a method called HDFL based on our observations, which is the first framework to our knowledge that comprehensively considers the multiple aforementioned important challenges of practical FL systems. We implement HDFL on a real distributed system and evaluate it on multiple benchmark datasets which show that HDFL achieves better Pareto frontier compared to both the state-of-the-practice and state-of-the-art systems with up to 4-10% better model accuracy, 33% improved good-intent fairness, 63% lower cost, and 17% faster training time.

In Federated Learning (FL), clients independently train local models and share them with a central aggregator to build a global model. Impermissibility to access clients' data and collaborative training make FL appealing for applications with data-privacy concerns, such as medical imaging. However, these FL characteristics pose unprecedented challenges for debugging. When a global model's performance deteriorates, identifying the responsible rounds and clients is a major pain point. Developers resort to trial-and-error debugging with subsets of clients, hoping to increase the global model's accuracy or let future FL rounds retune the model, which are time-consuming and costly. We design a systematic fault localization framework, Fedde-bug,that advances the FL debugging on two novel fronts. First, Feddebug enables interactive debugging of realtime collaborative training in FL by leveraging record and replay techniques to construct a simulation that mirrors live FL. Feddebug'sbreakpoint can help inspect an FL state (round, client, and global model) and move between rounds and clients' models seam-lessly, enabling a fine-grained step-by-step inspection. Second, Feddebug automatically identifies the client(s) responsible for lowering the global model's performance without any testing data and labels-both are essential for existing debugging techniques. Feddebug's strengths come from adapting differential testing in conjunction with neuron activations to determine the client(s) deviating from normal behavior. Feddebug achieves 100% accuracy in finding a single faulty client and 90.3% accuracy in finding multiple faulty clients. Feddebug's interactive de-bugging incurs 1.2% overhead during training, while it localizes a faulty client in only 2.1% of a round's training time. With FedDebug,we bring effective debugging practices to federated learning, improving the quality and productivity of FL application developers.

Serverless platforms offer on-demand computation and represent a significant shift from previous platforms that typically required resources to be pre-allocated (e.g., virtual machines). As serverless platforms have evolved, they have become suitable for a much wider range of applications than their original use cases. However, storage access remains a pain point that holds serverless back from becoming a completely generic computation platform.

Cloud object storage such as AWS S3 is cost-effective and highly elastic but relatively slow, while high-performance cloud storage such as AWS ElastiCache is expensive and provides limited elasticity. We present a new cloud storage service called ServerlessMemory, which stores data using the memory of serverless functions. ServerlessMemory employs a sliding-window-based memory management strategy inspired by the garbage collection mechanisms used in the programming language to effectively segregate hot/cold data and provides fine-grained elasticity, good performance, and a pay-per-access cost model with extremely low cost. We then design and implement InfiniStore, a persistent and elastic cloud storage system, which seamlessly couples the function-based ServerlessMemory layer with a persistent, inexpensive cloud object store layer. InfiniStore enables durability despite function failures using a fast parallel recovery scheme built on the auto-scaling functionality of a FaaS (Function-as-a-Service) platform. We evaluate InfiniStore extensively using both microbenchmarking and two real-world applications. Results show that InfiniStore has more performance benefits for objects larger than 10 MB compared to AWS ElastiCache and Anna, and InfiniStore achieves 26.25% and 97.24% tenant-side cost reduction compared to InfiniCache and ElastiCache, respectively.

In federated learning (FL), clients collectively train a global machine learning model with their own local data. Without sharing sensitive raw data, each client in FL only sends updated weights to consider privacy and security concerns. Most of existing FL works focus mainly on improving model accuracy and training time, but only a few works focus on FL incentive mechanisms. To build a high performance model after FL training, clients need to provide high quality and large amounts of data. However, in real FL scenarios, high-quality clients are reluctant to participate in FL process without reasonable compensation, because clients are self-interested and other clients can be business competitors. Even participation incurs some cost for contributing to the FL model with their local dataset. To address this problem, we propose TIFF, a novel tokenized incentive mechanism, where tokens are used as a means of paying for the services of providing participants and the training infrastructure. Without payment delays, participation can be monetized as both providers and consumers, which promotes continued long-term participation of high-quality data parties. Additionally, paid tokens are reimbursed to each client as consumers according to our newly proposed metrics (such as token reduction ratio and utility improvement ratio), which keeps clients engaged in FL process as consumers. To measure data quality, accuracy is calculated in training without additional overheads. We leverage historical accuracy records and random exploration to select high-utility participants and to prevent overfitting. Results show that TIFF provides more tokens to normal providers by up to 6.9% and less tokens to malicious providers by up to 18.1%, achieving improvement of the final model accuracy by up to 7.4%, compared to the default approach.

Federated learning has emerged as a privacy-preserving machine learning approach where multiple parties can train a single model without sharing their raw training data. Federated learning typically requires the utilization of multi-party computation techniques to provide strong privacy guarantees by ensuring that an untrusted or curious aggregator cannot obtain isolated replies from parties involved in the training process, thereby preventing potential inference attacks. Until recently, it was thought that some of these secure aggregation techniques were sufficient to fully protect against inference attacks coming from a curious aggregator. However, recent research has demonstrated that a curious aggregator can successfully launch a disaggregation attack to learn information about model updates of a target party. This paper presents DeTrust-FL, an efficient privacy-preserving federated learning framework for addressing the lack of transparency that enables isolation attacks, such as disaggregation attacks, during secure aggregation by assuring that parties’ model updates are included in the aggregated model in a private and secure manner. DeTrust-FL proposes a decentralized trust consensus mechanism and incorporates a recently proposed decentralized functional encryption scheme in which all parties agree on a participation matrix before collaboratively generating decryption key fragments, thereby gaining control and trust over the secure aggregation process in a decentralized setting. Our experimental evaluation demonstrates that DeTrust-FL outperforms state-of-the-art FE-based secure multi-party aggregation solutions in terms of training time and reduces the volume of data transferred. In contrast to existing approaches, this is achieved without creating any trust dependency on external trusted entities.

Federated learning (FL) is becoming an important distributed machine learning approach that considers privacy and security concerns while training a shared model across various clients with localized data. One of the key challenges in FL is heterogeneity in both hardware resources and local datasets due to the nature of incorporating diverse clients. Given the resource heterogeneity, the availability of participating clients is not stable over time and their resource usage patterns become dynamic. This leads to resource wastage and straggler issues. Additional challenges are introduced due to data heterogeneity, causing model biasness and poor model performance. However, most existing FL systems are not well suited to heterogeneous environments because those approaches are not adaptive to various and dynamically changing resource usage patterns and accuracy trends during training process. To this end, we propose a heterogeneity-aware scheduling which is adaptive to the accuracy trends and various resource usage patterns. Our proposed scheduling provides different scheduling knobs for achieving different goals such as resource-efficient fast training, resource fairness, accuracy fairness, and high model performance. To the best of our knowledge, this is the first effort to mitigate effects of resource and data heterogeneity while providing adaptive scheduling based on dynamically changing resource usage patterns and accuracy trends.

Federated learning (FL) facilitates the training and deploying AI models on edge devices. Preserving user data privacy in FL introduces several challenges, including expensive communication costs, limited resources, and data heterogeneity. In this paper, we propose SPATL, an FL method that addresses these issues by: (a) introducing a salient parameter selection agent and communicating selected parameters only; (b) splitting a model into a shared encoder and a local predictor, and transferring its knowledge to heterogeneous clients via the locally customized predictor. Additionally, we leverage a gradient control mechanism to further speed up model convergence and increase robustness of training processes. Experiments demonstrate that SPATL reduces communication overhead, accelerates model inference, and enables stable training processes with better results compared to state-of-the-art methods. Our approach reduces communication cost by up to 86.45%, accelerates local inference by reducing up to 39.7% FLOPs on VGG-11, and requires 7.4× less communication overhead when training ResNet-20.11Code is available at: https://github.com/yusx-swapp/SPATL

Federated learning has arisen as a mechanism to allow multiple participants to collaboratively train a model without sharing their data. In these settings, participants (workers) may not trust each other fully; for instance, a set of competitors may collaboratively train a machine learning model to detect fraud. The workers provide local gradients that a central server uses to update a global model. This global model can be corrupted when Byzantine workers send malicious gradients, which necessitates robust methods for aggregating gradients that mitigate the adverse effects of Byzantine inputs. Existing robust aggregation algorithms are often computationally expensive and only effective under strict assumptions. In this paper, we introduce LayerwisE Gradient AggregatTiOn (LEGATO), an aggregation algorithm that is, by contrast, scalable and generalizable. Informed by a study of layer-specific responses of gradients to Byzantine attacks, LEGATO employs a dynamic gradient reweighing scheme that is novel in its treatment of gradients based on layer-specific robustness. We show that LEGATO is more computationally efficient than multiple state-of-the-art techniques and more generally robust across a variety of attack settings in practice. We also demonstrate LEGATO's benefits for gradient descent convergence in the absence of an attack.

Data heterogeneity has been identified as one of the key features in federated learning but often overlooked in the lens of robustness to adversarial attacks. This paper focuses on characterizing and understanding its impact on backdooring attacks in federated learning through comprehensive experiments using synthetic and the LEAF benchmarks. The initial impression driven by our experimental results suggests that data heterogeneity is the dominant factor in the effectiveness of attacks and it may be a redemption for defending against backdooring as it makes the attack less efficient, more challenging to design effective attack strategies, and the attack result also becomes less predictable. However, with further investigations, we found data heterogeneity is more of a curse than a redemption as the attack effectiveness can be significantly boosted by simply adjusting the client-side backdooring timing. More importantly, data heterogeneity may result in overfitting at the local training of benign clients, which can be utilized by attackers to disguise themselves and fool skewed-feature based defenses. In addition, effective attack strategies can be made by adjusting attack data distribution. Finally, we discuss the potential directions of defending the curses brought by data heterogeneity. The results and lessons learned from our extensive experiments and analysis offer new insights for designing robust federated learning methods and systems.

Federated learning (FL) involves training a model over massive distributed devices, while keeping the training data localized and private. This form of collaborative learning exposes new tradeoffs among model convergence speed, model accuracy, balance across clients, and communication cost, with new challenges including: (1) straggler problem---where clients lag due to data or (computing and network) resource heterogeneity, and (2) communication bottleneck---where a large number of clients communicate their local updates to a central server and bottleneck the server. Many existing FL methods focus on optimizing along only one single dimension of the tradeoff space. Existing solutions use asynchronous model updating or tiering-based, synchronous mechanisms to tackle the straggler problem. However, asynchronous methods can easily create a communication bottleneck, while tiering may introduce biases that favor faster tiers with shorter response latencies.

Docker containers have become a prominent solution for supporting modern enterprise applications due to the highly desirable features of isolation, low overhead, and efficient packaging of the application’s execution environment. Containers are created from images which are shared between users via a registry. The amount of data registries store is massive. For example, Docker Hub, a popular public registry, stores at least half a million public images. In this article, we analyze over 167 TB of uncompressed Docker Hub images, characterize them using multiple metrics and evaluate the potential of file-level deduplication. Our analysis helps to make conscious decisions when designing storage for containers in general and Docker registries in particular. For example, only 3 percent of the files in images are unique while others are redundant file copies, which means file-level deduplication has a great potential to save storage space. Furthermore, we carry out a comprehensive analysis of both small I/O request performance and copy-on-write performance for multiple popular container storage drivers. Our findings can motivate and help improve the design of data reduction and caching methods for images, pulling optimizations for registries, and storage drivers.

Executing complex, burst-parallel, directed acyclic graph (DAG) jobs poses a major challenge for serverless execution frameworks, which will need to rapidly scale and schedule tasks at high throughput, while minimizing data movement across tasks. We demonstrate that, for serverless parallel computations, decentralized scheduling enables scheduling to be distributed across Lambda executors that can schedule tasks in parallel, and brings multiple benefits, including enhanced data locality, reduced network I/Os, automatic resource elasticity, and improved cost effectiveness. We describe the implementation and deployment of our new serverless parallel framework, called Wukong, on AWS Lambda. We show that Wukong achieves near-ideal scalability, executes parallel computation jobs up to 68.17X faster, reduces network I/O by multiple orders of magnitude, and achieves 92.96% tenant-side cost savings compared to numpywren.

Federated Learning (FL) enables learning a shared model acrossmany clients without violating the privacy requirements. One of the key attributes in FL is the heterogeneity that exists in both resource and data due to the differences in computation and communication capacity, as well as the quantity and content of data among different clients. We conduct a case study to show that heterogeneity in resource and data has a significant impact on training time and model accuracy in conventional FL systems. To this end, we propose TiFL, a Tier-based Federated Learning System, which divides clients into tiers based on their training performance and selects clients from the same tier in each training round to mitigate the straggler problem caused by heterogeneity in resource anddata quantity. To further tame the heterogeneity caused by non-IID (Independent and Identical Distribution) data and resources, TiFL employs an adaptive tier selection approach to update the tiering on-the-fly based on the observed training performance and accuracy. We prototype TiFL in a FL testbed following Google's FL architecture and evaluate it using the state-of-the-art FL benchmarks. Experimental evaluation shows that TiFL outperforms the conventional FL in various heterogeneous conditions. With the proposed adaptive tier selection policy, we demonstrate that TiFL achieves much faster training performance while achieving the same or better test accuracy across the board.

Enterprise KV stores are often not well suited for HPC applications, and thus cumbersome end-to-end KV design customization is required to meet the needs of modern HPC applications. To this end, in this article we present bespoKV, an adaptive, extensible, and scale-out KV store framework. bespoKV decouples the KV store design into the control plane for distributed management and the data plane for local data store. For the control plane, bespoKVprovides pre-built modules, called controlets, supporting common distributed functionalities (e.g., replication, consistency, and topology) and their various combinations. This decoupling allows bespoKV to take a user-provided single-server KV store, called a datalet, and transparently enables a scalable and fault-tolerant distributed KV store service. The resulting distributed stores are also adaptive to consistency or topology requirement changes and can be easily extended for new types of services. Such specializations enable innovative uses of KV stores in HPC applications, especially for emerging applications that utilize KV-friendly workloads. We evaluate bespoKV in a local testbed as well as in a public cloud settings. Experiments show that bespoKV-enabled distributed KV stores scale horizontally to a large number of nodes, and performs comparably and sometimes 1.2× to 2.6× better than the state-of-the-art systems.

Docker container images are typically stored in a centralized registry to allow easy sharing of images. However, with the growing popularity of containerized software, the number of images that a registry needs to store and the rate of requests it needs to serve are increasing rapidly. Current registry design requires hosting registry services across multiple loosely connected servers with different roles such as load balancers, proxies, registry servers, and object storage servers. Due to the various individual components, registries are hard to scale and benefits from optimizations such as caching are limited. In this paper we propose, implement, and evaluate BOLT-a new hyperconverged design for container registries. In BOLT, all registry servers are part of a tightly connected cluster and play the same consolidated role: each registry server caches images in its memory, stores images in its local storage, and provides computational resources to process client requests. The design employs a custom consistent hashing function to take advantage of the layered structure and addressing of images and to load balance requests across different servers. Our evaluation using real production workloads shows that BOLT outperforms the conventional registry design significantly and improves latency by an order of magnitude and throughput by up to 5x. Compared to state-of-the-art, BOLT can utilize cache space more efficiently and serve up to 35% more requests from its cache. Furthermore, BOLT scales linearly and recovers from failure recovery without significant performance degradation.

Due to their tight isolation, low overhead, and efficient packaging of the execution environment, Docker containers have become a prominent solution for deploying modern applications. Containers are created from images which are stored in a Docker registry. An image consists of a list of layers which can be shared among images. Docker registries store a large amount of images and with the increasing popularity of Docker, they continue to grow. For example, Docker Hub-a popular public registry-stores more than half a million public images. In this paper, we analyze over 167TB of uncompressed Docker images and evaluate the potential of file-level deduplication in the registry. Our analysis reveals that only 3% of the files in images are unique and Docker's existing layer sharing mechanism is not sufficient to eliminate this profound redundancy. We then present the design of Slimmer-a Docker registry with file deduplication support-and conduct a simulation-based analysis of its performance implications.

Docker containers have become a prominent solution for supporting modern enterprise applications due to the highly desirable features of isolation, low overhead, and efficient packaging of the execution environment. Containers are created from images which are shared between users via a Docker registry. The amount of data Docker registries store is massive; for example, Docker Hub, a popular public registry, stores at least half a million public images. In this paper, we analyze over 167 TB of uncompressed Docker Hub images, characterize them using multiple metrics and evaluate the potential of file-level deduplication in Docker Hub. Our analysis helps to make conscious decisions when designing storage for containers in general and Docker registries in particular. For example, only 3% of the files in images are unique, which means file-level deduplication has a great potential to save storage space for the registry. Our findings can motivate and help improve the design of data reduction, caching, and pulling optimizations for registries.

Exponential increases in complexity and scale make variability a growing threat to sustaining HPC performance at exascale. Performance variability in HPC I/O is common, acute, and formidable. We take the first step towards comprehensively studying linear and nonlinear approaches to modeling HPC I/O system variability in an effort to demonstrate that variability is often a predictable artifact of system design. Using over 8 months of data collection on 6 identical systems, we propose and validate a modeling and analysis approach (MOANA) that predicts HPC I/O variability for thousands of software and hardware configurations on highly parallel shared-memory systems. Our findings indicate nonlinear approaches to I/O variability prediction are an order of magnitude more accurate than linear regression techniques. We demonstrate the use of MOANA to accurately predict the confidence intervals of unmeasured I/O system configurations for a given number of repeat runs - enabling users to quantitatively balance experiment duration with statistical confidence.

Warehouse-scale cloud datacenters co-locate workloads with different and often complementary characteristics for improved resource utilization. To better understand the challenges in managing such intricate, heterogeneous workloads while providing quality-assured resource orchestration and user experience, we analyze Alibaba's co-located workload trace, the first publicly available dataset with precise information about the category of each job. Two types of workload-long-running, user-facing, containerized production jobs, and transient, highly dynamic, non-containerized, and non-production batch jobs- are running on a shared cluster of 1313 machines. Through workload characterization, we find evidences that imply that one workload scheduler makes seemingly independent scheduling decisions regardless of the co-existence of the other. This upsurges an imminent need for a more integrated, global coordinating system that transparently connect multiple resource schedulers together and cohesively coordinates the multiple heterogeneous workloads for greater efficiency. Our multifaceted analysis reveals insights that we believe are useful for system designers and IT practitioners working on cluster management systems.

NAND flash-based Solid State Devices (SSDs) offer the desirable features of high performance, energy efficiency, and fast growing capacity. Thus, the use of SSDs is increasing in distributed storage systems. A key obstacle in this context is that the natural unbalance in distributed I/O workloads can result in wear imbalance across the SSDs in a distributed setting. This, in turn can have significant impact on the reliability, performance, and lifetime of the storage deployment. Extant load balancers for storage systems do not consider SSD wear imbalance when placing data, as the main design goal of such balancers is to extract higher performance. Consequently, data migration is the only common technique for tackling wear imbalance, where existing data is moved from highly loaded servers to the least loaded ones. In this paper, we explore an innovative holistic approach, Chameleon, that employs data redundancy techniques such as replication and erasure-coding, coupled with endurance-aware write offloading, to mitigate wear level imbalance in distributed SSD-based storage. Chameleon aims to balance the wear among different flash servers while meeting desirable objectives of: extending life of flash servers; improving I/O performance; and avoiding bottlenecks. Evaluation with a 50 node SSD cluster shows that Chameleon reduces the wear distribution deviation by 81% while improving the write performance by up to 33%.

The use of cloud object stores has been growing rapidly in recent years as they combine key advantages such as HTTP-based RESTful APIs, high availability, elasticity with a "pay-as-you-go" pricing model that allows applications to scale as needed. The current practice is to either use a single set of configuration parameters or rely on statically configured storage policies for a cloud object store deployment, even when the store is used to support different types of applications with evolving requirements. This crucial mismatch between the different applications requirements and capabilities of the object store is problematic and should be addressed to achieve high efficiency and performance.

As the cloud services journey through their lifecycle towards becoming commodities, the demand is increasing for "pay-per-use" pricing model. In this model, users are charged for the amount of resources, e.g., Volume of transactions, CPU usage, etc., being consumed during a given time period. Software as a Service (SaaS) providers charging their customers via pay-per-use (e.g., Microsoft Azure Web Services) and facing Infrastructure as a Service (IaaS) costs per VM per month (e.g., Soft Layer) have to carefully choose and scale their non-revenue generating service management infrastructure to penetrate and stay in the market. In this paper, we focus on the metering and rating aspects of cloud service management, and their scalability with the SaaS business and operational changes. We design a framework for cloud service providers to scale their revenue management systems in a cost-aware manner, where the deployment of these revenue systems dynamically uses existing or newly provisioned SaaS VMs, instead of the extant approach of using dedicated setups. Our experimental analysis shows that service management related tasks can be offloaded to the existing VMs with at most 15% overhead in CPU utilization, 10% overhead for memory usage, and negligible overhead for I/O and network usage. We used traces from IBM production servers to mimic the load on VMs. By dynamically scaling the service management setup, we were able to adapt to increasing metering data processing requirements without incurring additional cost, while preserving the infrastructure footprint.

As the cloud services journey through their life-cycle towards commodities, cloud service providers have to carefully choose the metering and rating tools and scale their infrastructure to effectively process the collected metering data. In this paper, we focus on the metering and rating aspects of the revenue management and their adaptability to business and operational changes. We design a framework for IT cloud service providers to scale their revenue systems in a cost-aware manner. The main idea is to dynamically use existing or newly provisioned SaaS VMs, instead of dedicated setups, for deploying the revenue management systems. At on-boarding of new customers, our framework performs off-line analysis to recommend appropriate revenue tools and their scalable distribution by predicting the need for resources based on historical usage. This allows the revenue management to adapt to the ever evolving business context. We evaluated our framework on a test bed of 20 physical machines that were used to deploy 12 VMs within Open Stack environment. Our analysis shows that service management related tasks can be offloaded to the existing VMs with at most 15% overhead in CPU utilization, 10% overhead for memory usage, and negligible overhead for I/O and network usage. By dynamically scaling the setup, we were able to reduce the metering data processing time by many folds without incurring any additional cost.

The need for novel data analysis is urgent in the face of a data deluge from modern applications. Traditional approaches to data analysis incur significant data movement costs, moving data back and forth between the storage system and the processor. Emerging Active Flash devices enable processing on the flash, where the data already resides. An array of such Active Flash devices allows us to revisit how analysis workflows interact with storage systems. By seamlessly blending together the flash storage and data analysis, we create an analysis workflow-aware storage system, AnalyzeThis. Our guiding principle is that analysis-awareness be deeply ingrained in each and every layer of the storage, elevating data analyses as first-class citizens, and transforming AnalyzeThis into a potent analytics-aware appliance. We implement the AnalyzeThis storage system atop an emulation platform of the Active Flash array. Our results indicate that AnalyzeThis is viable, expediting workflow execution and minimizing data movement.

Hadoop has become the de-facto large-scale data processing framework for modern analytics applications. A major obstacle for sustaining high performance and scalability in Hadoop is managing the data growth while meeting the ever higher I/O demand. To this end, a promising trend in storage systems is to utilize hybrid and heterogeneous devices - Solid State Disks (SSD), ram disks and Network Attached Storage (NAS), which can help achieve very high I/O rates at acceptable cost. However, the Hadoop Distributed File System (HDFS) that is unable to exploit such heterogeneous storage. This is because HDFS works on the assumption that the underlying devices are homogeneous storage blocks, disregarding their individual I/O characteristics, which leads to performance degradation. In this paper, we present hatS, a Heterogeneity-Aware Tiered Storage, which is a novel redesign of HDFS into a multi-tiered storage system that seamlessly integrates heterogeneous storage technologies into the Hadoop ecosystem. hatS also proposes data placement and retrieval policies, which improve the utilization of the storage devices based on their characteristics such as I/O throughput and capacity. We evaluate hatS using an actual implementation on a medium-sized cluster consisting of HDDs and two types of SSDs (i.e., SATA SSD and PCIe SSD). Experiments show that hatS achieves 32.6% higher read bandwidth, on average, than HDFS for the test Hadoop jobs (such as Grep and Test DFSIO) by directing 64% of the I/O accesses to the SSD tiers. We also evaluate our approach with trace-driven simulations using synthetic Facebook workloads, and show that compared to the standard setup, hatS improves the average I/O rate by 36%, which results in 26% improvement in the job completion time.

The use of economical, low-power microservers comprising of embedded CPUs is on the rise in supporting a myriad of applications. State of the art microservers can already match the performance of low-end traditional servers, and have been advocated as an energy-efficient alternative computing substrate for data centers as well. In this paper, we explore whether cluster comprising microservers can support the popular Hadoop platform. We conduct a quantitative study of six representative Hadoop applications on five hardware configurations. To compare the different clusters, we also define a comprehensive metric, PerfEC, which unifies the performance, energy consumption, and the acquisition and operating costs of the applications, and helps identify appropriate clusters for Hadoop applications. Experiments on our test clusters suggest that for applications such as TeraSort, RandomWriter and Grep microservers offer up to two orders of magnitude better efficiency in terms of PerfEC than traditional clusters. Similarly, a 3000-node cluster simulation driven by a real-world trace from Facebook shows that on average the studied microservers can match the performance of standard servers, while providing up to 31% energy savings at only 60% of the acquisition cost. We also compare PerfEC to the extant Total Cost of Ownership (TCO) metric, and find that our approach is better able to capture the trade-offs involved.

Enterprise Hadoop applications now routinely comprise complex workflows that are managed by specialized workflow schedulers such as Oozie. The resources are assumed to be similar or homogeneous and data locality is often the only scheduling constraint considered. However, introduction of specialized architectures and regular system upgrades lead to Hadoop data center hardware becoming increasingly heterogeneous, in that a data center may have several clusters each boasting different characteristics. However, the workflow scheduler is not aware of such heterogeneity, and thus cannot ensure that a cluster selected based on data locality is also suitable for supporting the jobs efficiently in terms of execution time and resource consumption. In this paper, we adopt a quantitative approach where we first study detailed behavior of various representative Hadoop applications running on four different hardware configurations. Next, we incorporate this information into a hardware-aware scheduler, ØSched, to improve the resource application match. To ensure that job associated data is available locally (or nearby) to a cluster in a multi-cluster deployment, we configure a single Hadoop Distributed File System (HDFS) instance across all the participating clusters. We also design and implement region-aware data placement and retrieval for HDFS in order to reduce the network overhead and achieve cluster-level data locality. We evaluate our approach using experiments on Amazon EC2 with four clusters of eight homogeneous nodes each, where each cluster has a different hardware configuration. We find that ØSched's optimized placement of applications across the test clusters reduces the execution time of the test applications by 18.7%, on average, when compared to extant hardware oblivious scheduling. Moreover, our HDFS enhancement increases the I/O throughput by up to 23% and the average I/O rate by up to 26% for the TestDFSIO benchmark.