Search

Modern artificial intelligence (AI) systems are powered by foundation models. This paper presents a new set of foundation models, called Llama 3. It is a herd of language models that natively support multilinguality, coding, reasoning, and tool usage. Our largest model is a dense Transformer with 405B parameters and a context window of up to 128K tokens. This paper presents an extensive empirical evaluation of Llama 3. We find that Llama 3 delivers comparable quality to leading language models such as GPT-4 on a plethora of tasks. We publicly release Llama 3, including pre-trained and post-trained versions of the 405B parameter language model and our Llama Guard 3 model for input and output safety. The paper also presents the results of experiments in which we integrate image, video, and speech capabilities into Llama 3 via a compositional approach. We observe this approach performs competitively with the state-of-the-art on image, video, and speech recognition tasks. The resulting models are not yet being broadly released as they are still under development.

Palimpsests refer to historical manuscripts where erased writings have been partially covered by the superimposition of a second writing. By employing imaging techniques, e.g., multispectral imaging, it becomes possible to identify features that are imperceptible to the naked eye, including faded and erased inks. When dealing with overlapping inks, Artificial Intelligence techniques can be utilized to disentangle complex nodes of overlapping letters. In this work, we propose deep learning-based semantic segmentation as a method for identifying and segmenting individual letters in overlapping characters. The experiment was conceived as a proof of concept, focusing on the palimpsests of the Ars Grammatica by Prisciano as a case study. Furthermore, caveats and prospects of our approach combined with multispectral imaging are also discussed.

Internet access is a special resource of which needs has become universal across the public whereas the service is operated in the private sector. Mobile Network Operators (MNOs) put efforts for management, planning, and optimization; however, they do not link such activities to socioeconomic fairness. In this paper, we make a first step towards understanding the relation between socioeconomic status of customers and network performance, and investigate potential discrimination in network deployment and management. The scope of our study spans various aspects, including urban geography, network resource deployment, data consumption, and device distribution. A novel methodology that enables a geo-socioeconomic perspective to mobile network is developed for the study. The results are based on an actual infrastructure in multiple cities, covering millions of users densely covering the socioeconomic scale. We report a thorough examination of the fairness status, its relationship with various structural factors, and potential class specific solutions.

Autonomous driving services rely heavily on sensors such as cameras, LiDAR, radar, and communication modules. A common practice of processing the sensed data is using a high-performance computing unit placed inside the vehicle, which deploys AI models and algorithms to act as the brain or administrator of the vehicle. The vehicular data generated from average hours of driving can be up to 20 Terabytes depending on the data rate and specification of the sensors. Given the scale and fast growth of services for autonomous driving, it is essential to improve the overall energy and environmental efficiency, especially in the trend towards vehicular electrification (e.g., battery-powered). Although the areas have seen significant advancements in sensor technologies, wireless communications, computing and AI/ML algorithms, the challenge still exists in how to apply and integrate those technology innovations to achieve energy efficiency. This survey reviews and compares the connected vehicular applications, vehicular communications, approximation and Edge AI techniques. The focus is on energy efficiency by covering newly proposed approximation and enabling frameworks. To the best of our knowledge, this survey is the first to review the latest approximate Edge AI frameworks and publicly available datasets in energy-efficient autonomous driving. The insights and vision from this survey can be beneficial for the collaborative driving service development on low-power and memory-constrained systems and also for the energy optimization of autonomous vehicles.

Distributed (or Federated) learning enables users to train machine learning models on their very own devices, while they share only the gradients of their models usually in a differentially private way (utility loss). Although such a strategy provides better privacy guarantees than the traditional centralized approach, it requires users to blindly trust a centralized infrastructure that may also become a bottleneck with the increasing number of users. In this paper, we design and implement P4L: a privacy preserving peer-to-peer learning system for users to participate in an asynchronous, collaborative learning scheme without requiring any sort of infrastructure or relying on differential privacy. Our design uses strong cryptographic primitives to preserve both the confidentiality and utility of the shared gradients, a set of peer-to-peer mechanisms for fault tolerance and user churn, proximity and cross device communications. Extensive simulations under different network settings and ML scenarios for three real-life datasets show that P4L provides competitive performance to baselines, while it is resilient to different poisoning attacks. We implement P4L and experimental results show that the performance overhead and power consumption is minimal (less than 3mAh of discharge).

Many real-time applications (e.g., Augmented/Virtual Reality, cognitive assistance) rely on Deep Neural Networks (DNNs) to process inference tasks. Edge computing is considered a key infrastructure to deploy such applications, as moving computation close to the data sources enables us to meet stringent latency and throughput requirements. However, the constrained nature of edge networks poses several additional challenges to the management of inference workloads: edge clusters can not provide unlimited processing power to DNN models, and often a trade-off between network and processing time should be considered when it comes to end-to-end delay requirements. In this paper, we focus on the problem of scheduling inference queries on DNN models in edge networks at short timescales (i.e., few milliseconds). By means of simulations, we analyze several policies in the realistic network settings and workloads of a large ISP, highlighting the need for a dynamic scheduling policy that can adapt to network conditions and workloads. We therefore design ASET, a Reinforcement Learning based scheduling algorithm able to adapt its decisions according to the system conditions. Our results show that ASET effectively provides the best performance compared to static policies when scheduling over a distributed pool of edge resources.

Federated Learning (FL) has recently emerged as a popular solution to distributedly train a model on user devices improving user privacy and system scalability. Major Internet companies have deployed FL in their applications for specific use cases (e.g., keyboard prediction or acoustic keyword trigger), and the research community has devoted significant attention to improving different aspects of FL (e.g., accuracy, privacy, efficiency). However, there is still a lack of a practical system to enable easy collaborative cross-silo FL training, in the context of mobile environments. In this work, we bridge this gap and propose FLaME, an end-to-end system (i.e., client-side framework and libraries, and central server) to enable intra- and inter-app training on mobile devices with different types of IID and NonIID data distributions, in a secure and easy to deploy fashion. Our design solves major technical challenges such as on-device training, secure and private single and cross-app model training, while being offered in an "as a service" model. We implement FLaME for Android devices and experimentally evaluate its performance in-lab and in-wild, on more than 140 users for over a month. Our results show the feasibility and benefits of the design in a realistic mobile context and provide several insights to the FL community on the practicality and usage of FL in the wild., Comment: 12 pages, 6 figures, 46 references

Federated learning (FL), where data remains at the federated clients, and where only gradient updates are shared with a central aggregator, was assumed to be private. Recent work demonstrates that adversaries with gradient-level access can mount successful inference and reconstruction attacks. In such settings, differentially private (DP) learning is known to provide resilience. However, approaches used in the status quo (\ie central and local DP) introduce disparate utility vs. privacy trade-offs. In this work, we take the first step towards mitigating such trade-offs through {\em hierarchical FL (HFL)}. We demonstrate that by the introduction of a new intermediary level where calibrated DP noise can be added, better privacy vs. utility trade-offs can be obtained; we term this {\em hierarchical DP (HDP)}. Our experiments with 3 different datasets (commonly used as benchmarks for FL) suggest that HDP produces models as accurate as those obtained using central DP, where noise is added at a central aggregator. Such an approach also provides comparable benefit against inference adversaries as in the local DP case, where noise is added at the federated clients.

The proliferation of emergent network applications (e.g., telesurgery, metaverse) is increasing the difficulty of managing modern communication networks. These applications entail stringent network requirements (e.g., ultra-low deterministic latency), which hinders network operators to manage their resources efficiently. In this article, we introduce the network digital twin (NDT), a renovated concept of classical network modeling tools whose goal is to build accurate data-driven network models that can operate in real-time. We describe the general architecture of the NDT and argue that modern machine learning (ML) technologies enable building some of its core components. Then, we present a case study that leverages a ML-based NDT for network performance evaluation and apply it to routing optimization in a QoS-aware use case. Lastly, we describe some key open challenges and research opportunities yet to be explored to achieve effective deployment of NDTs in real-world networks., Comment: 7 pages, 4 figures. arXiv admin note: text overlap with arXiv:2201.01144

Pose estimation of an uncooperative space resident object is a key asset towards autonomy in close proximity operations. In this context monocular cameras are a valuable solution because of their low system requirements. However, the associated image processing algorithms are either too computationally expensive for real time on-board implementation, or not enough accurate. In this paper we propose a pose estimation software exploiting neural network architectures which can be scaled to different accuracy-latency trade-offs. We designed our pipeline to be compatible with Edge Tensor Processing Units to show how low power machine learning accelerators could enable Artificial Intelligence exploitation in space. The neural networks were tested both on the benchmark Spacecraft Pose Estimation Dataset, and on the purposely developed Cosmo Photorealistic Dataset, which depicts a COSMO-SkyMed satellite in a variety of random poses and steerable solar panels orientations. The lightest version of our architecture achieves state-of-the-art accuracy on both datasets but at a fraction of networks complexity, running at 7.7 frames per second on a Coral Dev Board Mini consuming just 2.2W., Comment: Improved literature review; added Figure 2; revised tables for better readibility; corrected typos

The proliferation of emergent network applications (e.g., AR/VR, telesurgery, real-time communications) is increasing the difficulty of managing modern communication networks. These applications typically have stringent requirements (e.g., ultra-low deterministic latency), making it more difficult for network operators to manage their network resources efficiently. In this article, we propose the Digital Twin Network (DTN) as a key enabler for efficient network management in modern networks. We describe the general architecture of the DTN and argue that recent trends in Machine Learning (ML) enable building a DTN that efficiently and accurately mimics real-world networks. In addition, we explore the main ML technologies that enable developing the components of the DTN architecture. Finally, we describe the open challenges that the research community has to address in the upcoming years in order to enable the deployment of the DTN in real-world scenarios., Comment: 7 pages, 4 figures

We test the hypothesis that protest participation decisions in an adult population of potential climate protesters are interdependent. Subjects (n=1,510) from the four largest German cities were recruited two weeks before protest date. We measured participation (ex post) and beliefs about the other subjects' participation (ex ante) in an online survey, used a randomized informational intervention to induce exogenous variance in beliefs, and estimated the causal effect of a change in belief on the probability of participation using a control function approach. Participation decisions are found to be strategic substitutes: a one percentage-point increase of belief causes a .67 percentage-point decrease in the probability of participation in the average subject., Comment: 21 pages, 6 pages appendix, 4 figures

Serverless Computing is a virtualisation-related paradigm that promises to simplify application management and to solve the last challenges in the field: scale down and easy to use. The implied cost reduction, coupled with a simplified management of underlying applications, are expected to further push the adoption of virtualisation-based solutions, including cloud-computing or telco-cloud solutions. However, in this quest for efficiency, security is not ranked among the top priorities, also because of the (misleading) belief that current solutions developed for virtualised environments could be applied (as is) to this new paradigm. Unfortunately, this is not the case, due to the highlighted idiosyncratic features of serverless computing. In this paper, we review the current serverless architectures, abstract and categorise their founding principles, and provide an in depth analyse of them from the point of view of security, referring to principles and practices of the cybersecurity domain. In particular, we show the security shortcomings of the analysed serverless architectural paradigms, point to possible countermeasures, and highlight a few research directions.

Streaming 360{\deg} video demands high bandwidth and low latency, and poses significant challenges to Internet Service Providers (ISPs) and Mobile Network Operators (MNOs). The identification of 360{\deg} video traffic can therefore benefits fixed and mobile carriers to optimize their network and provide better Quality of Experience (QoE) to the user. However, end-to-end encryption of network traffic has obstructed identifying those 360{\deg} videos from regular videos. As a solution this paper presents 360NorVic, a near-realtime and offline Machine Learning (ML) classification engine to distinguish 360{\deg} videos from regular videos when streamed from mobile devices. We collect packet and flow level data for over 800 video traces from YouTube & Facebook accounting for 200 unique videos under varying streaming conditions. Our results show that for near-realtime and offline classification at packet level, average accuracy exceeds 95%, and that for flow level, 360NorVic achieves more than 92% average accuracy. Finally, we pilot our solution in the commercial network of a large MNO showing the feasibility and effectiveness of 360NorVic in production settings., Comment: 7 pages, 15 figures, accepted in Workshop on Network and OperatingSystem Support for Digital Audio and Video (NOSSDAV 21)

We propose and implement a Privacy-preserving Federated Learning ($PPFL$) framework for mobile systems to limit privacy leakages in federated learning. Leveraging the widespread presence of Trusted Execution Environments (TEEs) in high-end and mobile devices, we utilize TEEs on clients for local training, and on servers for secure aggregation, so that model/gradient updates are hidden from adversaries. Challenged by the limited memory size of current TEEs, we leverage greedy layer-wise training to train each model's layer inside the trusted area until its convergence. The performance evaluation of our implementation shows that $PPFL$ can significantly improve privacy while incurring small system overheads at the client-side. In particular, $PPFL$ can successfully defend the trained model against data reconstruction, property inference, and membership inference attacks. Furthermore, it can achieve comparable model utility with fewer communication rounds (0.54$\times$) and a similar amount of network traffic (1.002$\times$) compared to the standard federated learning of a complete model. This is achieved while only introducing up to ~15% CPU time, ~18% memory usage, and ~21% energy consumption overhead in $PPFL$'s client-side., Comment: 15 pages, 8 figures, accepted to MobiSys 2021

The closed design of mobile devices -- with the increased security and consistent user interfaces -- is in large part responsible for their becoming the dominant platform for accessing the Internet. These benefits, however, are not without a cost. Their operation of mobile devices and their apps is not easy to understand by either users or operators. We argue for recovering transparency and control on mobile devices through an extensible platform that can intercept and modify traffic before leaving the device or, on arrival, before it reaches the operating system. Conceptually, this is the same view of the traffic that a traditional middlebox would have at the far end of the first link in the network path. We call this platform ``middlebox zero'' or MBZ. By being on-board, MBZ also leverages local context as it processes the traffic and complements the network-wide view of standard middleboxes. We discuss the challenges of the MBZ approach, sketch a working design, and illustrate its potential with some concrete examples., Comment: The paper is a position piece under review