Federated Learning (FL) has emerged as a promising approach to address data privacy and confidentiality concerns by allowing multiple participants to construct a shared model without centralizing sensitive data. However, this decentralized paradigm introduces new security challenges, necessitating a comprehensive identification and classification of potential risks to ensure FL's security guarantees. This paper presents a comprehensive taxonomy of security and privacy challenges in Federated Learning (FL) across various machine learning models, including large language models. We specifically categorize attacks performed by the aggregator and participants, focusing on poisoning attacks, backdoor attacks, membership inference attacks, generative adversarial network (GAN) based attacks, and differential privacy attacks. Additionally, we propose new directions for future research, seeking innovative solutions to fortify FL systems against emerging security risks and uphold sensitive data confidentiality in distributed learning environments.
相關內容
Estimating the head pose of a person is a crucial problem for numerous applications that is yet mainly addressed as a subtask of frontal pose prediction. We present a novel method for unconstrained end-to-end head pose estimation to tackle the challenging task of full range of orientation head pose prediction. We address the issue of ambiguous rotation labels by introducing the rotation matrix formalism for our ground truth data and propose a continuous 6D rotation matrix representation for efficient and robust direct regression. This allows to efficiently learn full rotation appearance and to overcome the limitations of the current state-of-the-art. Together with new accumulated training data that provides full head pose rotation data and a geodesic loss approach for stable learning, we design an advanced model that is able to predict an extended range of head orientations. An extensive evaluation on public datasets demonstrates that our method significantly outperforms other state-of-the-art methods in an efficient and robust manner, while its advanced prediction range allows the expansion of the application area. We open-source our training and testing code along with our trained models: //github.com/thohemp/6DRepNet360.
In the presence of heterogeneous data, where randomly rotated objects fall into multiple underlying categories, it is challenging to simultaneously classify them into clusters and synchronize them based on pairwise relations. This gives rise to the joint problem of community detection and synchronization. We propose a series of semidefinite relaxations, and prove their exact recovery when extending the celebrated stochastic block model to this new setting where both rotations and cluster identities are to be determined. Numerical experiments demonstrate the efficacy of our proposed algorithms and confirm our theoretical result which indicates a sharp phase transition for exact recovery.
We consider the problem of checking the differential privacy of online randomized algorithms that process a stream of inputs and produce outputs corresponding to each input. This paper generalizes an automaton model called DiP automata (See arXiv:2104.14519) to describe such algorithms by allowing multiple real-valued storage variables. A DiP automaton is a parametric automaton whose behavior depends on the privacy budget $\epsilon$. An automaton $A$ will be said to be differentially private if, for some $\mathfrak{D}$, the automaton is $\mathfrak{D}\epsilon$-differentially private for all values of $\epsilon>0$. We identify a precise characterization of the class of all differentially private DiP automata. We show that the problem of determining if a given DiP automaton belongs to this class is PSPACE-complete. Our PSPACE algorithm also computes a value for $\mathfrak{D}$ when the given automaton is differentially private. The algorithm has been implemented, and experiments demonstrating its effectiveness are presented.
Analyzing memes on the internet has emerged as a crucial endeavor due to the impact this multi-modal form of content wields in shaping online discourse. Memes have become a powerful tool for expressing emotions and sentiments, possibly even spreading hate and misinformation, through humor and sarcasm. In this paper, we present the overview of the Memotion 3 shared task, as part of the DeFactify 2 workshop at AAAI-23. The task released an annotated dataset of Hindi-English code-mixed memes based on their Sentiment (Task A), Emotion (Task B), and Emotion intensity (Task C). Each of these is defined as an individual task and the participants are ranked separately for each task. Over 50 teams registered for the shared task and 5 made final submissions to the test set of the Memotion 3 dataset. CLIP, BERT modifications, ViT etc. were the most popular models among the participants along with approaches such as Student-Teacher model, Fusion, and Ensembling. The best final F1 score for Task A is 34.41, Task B is 79.77 and Task C is 59.82.
Every constructive model of computation (CMC) has an underlying composition mechanism for combining simple computation devices into more complex ones. Composition can be done by (explicitly or implicitly) defining control flow, data flow or any combination thereof. Control flow specifies the order in which individual computation devices are activated, whereas data flow defines how data is exchanged among them. Unfortunately, traditional CMCs either mix data and control or only consider one dimension explicitly, which makes it difficult to reason about data flow and control flow separately. Reasoning about these dimensions orthogonally is a crucial desideratum for optimisation, maintainability and verification purposes. In this paper, we introduce a novel model that explicitly treats data flow and control flow as separate dimensions, while providing modularity. As the model is rooted in category theory, it provides category-theoretic operations for compositionally constructing sequential or parallel composites. Compositionality entails that a composite exhibits the same properties as its respective constituents, including separation of concerns and modularity.
The introduction and advancements in Local Differential Privacy (LDP) variants have become a cornerstone in addressing the privacy concerns associated with the vast data produced by smart devices, which forms the foundation for data-driven decision-making in crowdsensing. While harnessing the power of these immense data sets can offer valuable insights, it simultaneously poses significant privacy risks for the users involved. LDP, a distinguished privacy model with a decentralized architecture, stands out for its capability to offer robust privacy assurances for individual users during data collection and analysis. The essence of LDP is its method of locally perturbing each user's data on the client-side before transmission to the server-side, safeguarding against potential privacy breaches at both ends. This article offers an in-depth exploration of LDP, emphasizing its models, its myriad variants, and the foundational structure of LDP algorithms.
Timely sampling and fresh information delivery are important in 6G communications. This is achieved by encoding samples into short packets/codewords for transmission, with potential decoding errors. We consider a broadcasting base station (BS) that samples information from multiple sources and transmits to respective destinations/users, using short-blocklength cyclic and deep learning (DL) based codes for error correction, and cyclic-redundancy-check (CRC) codes for error detection. We use a metric called reported age of information (AoI), abbreviated as RAoI, to measure the freshness of information, which increases from an initial value if the CRC reports a failure, else is reset. We minimize long-term average expected RAoI, subject to constraints on transmission power and distortion, for which we obtain age-agnostic randomized and age-aware drift-plus-penalty policies that decide which user to transmit to, with what message-word length and transmit power, and derive bounds on their performance. Simulations show that longer CRC codes lead to higher RAoI, but the RAoI achieved is closer to the true, genie-aided AoI. DL-based codes achieve lower RAoI. Finally, we conclude that prior AoI optimization literature with finite blocklengths substantially underestimates AoI because they assume that all errors can be detected perfectly without using CRC.
Large Language Models (LLMs) have shown excellent generalization capabilities that have led to the development of numerous models. These models propose various new architectures, tweaking existing architectures with refined training strategies, increasing context length, using high-quality training data, and increasing training time to outperform baselines. Analyzing new developments is crucial for identifying changes that enhance training stability and improve generalization in LLMs. This survey paper comprehensively analyses the LLMs architectures and their categorization, training strategies, training datasets, and performance evaluations and discusses future research directions. Moreover, the paper also discusses the basic building blocks and concepts behind LLMs, followed by a complete overview of LLMs, including their important features and functions. Finally, the paper summarizes significant findings from LLM research and consolidates essential architectural and training strategies for developing advanced LLMs. Given the continuous advancements in LLMs, we intend to regularly update this paper by incorporating new sections and featuring the latest LLM models.
Recently, Mutual Information (MI) has attracted attention in bounding the generalization error of Deep Neural Networks (DNNs). However, it is intractable to accurately estimate the MI in DNNs, thus most previous works have to relax the MI bound, which in turn weakens the information theoretic explanation for generalization. To address the limitation, this paper introduces a probabilistic representation of DNNs for accurately estimating the MI. Leveraging the proposed MI estimator, we validate the information theoretic explanation for generalization, and derive a tighter generalization bound than the state-of-the-art relaxations.
Residual networks (ResNets) have displayed impressive results in pattern recognition and, recently, have garnered considerable theoretical interest due to a perceived link with neural ordinary differential equations (neural ODEs). This link relies on the convergence of network weights to a smooth function as the number of layers increases. We investigate the properties of weights trained by stochastic gradient descent and their scaling with network depth through detailed numerical experiments. We observe the existence of scaling regimes markedly different from those assumed in neural ODE literature. Depending on certain features of the network architecture, such as the smoothness of the activation function, one may obtain an alternative ODE limit, a stochastic differential equation or neither of these. These findings cast doubts on the validity of the neural ODE model as an adequate asymptotic description of deep ResNets and point to an alternative class of differential equations as a better description of the deep network limit.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191