Recently, transformer-based large language models (LLMs) have become mainstream, particularly when used as agents. However, users with exceedingly sensitive data cannot benefit due to a lack of high-performance devices to train LLMs or limited access to large companies' LLM APIs. Secure two-party computing (2PC), especially the recently popular function secret sharing (FSS), enables secure inference of LLMs by protecting both users' inputs and LLM parameters from leakage. In this paper, we propose${\sf FLUTE}$, the first FSS-based 2PC inference framework for LLMs with partial transformer encryption. We first identify a subset of transformer core blocks by simulating an adversary$\mathcal {A}$attempting to recover model parameters layer by layer, and then design GPU-friendly FSS protocols for each module within the core blocks, optimizing communication using the matrix multiplication protocol of${\sf ABY2.0}$. Security analysis shows that our partial encryption scheme provides security comparable to encrypting the entire LLM, thereby enhancing the performance-security trade-off of the entire end-to-end secure inference. The experimental results in the latest LLM, Llama 3.1-8B, show that${\sf FLUTE}$outperforms the SOTA (SIGMA) by$3\times$in both latency and communication, and it achieves even greater advantages in larger LLMs, such as Llama 3.1-70B.

Camera-based Semantic Scene Completion (SSC) aims to infer the geometric structure and semantic information in the entire 3D scene from limited 2D images. However, due to the lack of geometric information in the image, existing methods tend to generate fuzzy completion and incorrect semantic boundaries. In this paper, we propose cross-modal knowledge distillation to address this issue, namely PI-Net, which guides the camera-based model to learn accurate 3D geometry to compensate for spatial surroundings information during training. Specifically, we propose a point cloud occupancy prediction model as the teacher, leveraging its output for strong depth supervision signals and spatial voxel information to enhance the student model. To facilitate effective distillation, we design depth guidance distillation to improve geometric predictions, and spatial guidance distillation to assist the student model in better capturing the structural information of the surrounding environment. Finally, prediction domain distillation is incorporated to facilitate holistic learning from point cloud to image. Experimental results demonstrate that PI-Net outperforms state-of-the-art camera-based methods on challenging benchmarks—SemanticKITTI and SSCBench-KITTI-360.

Federated learning is a collaborative training paradigm designed to protect private data and is widely used in the cooperative training of Internet of Things (IoT) devices. However, despite its focus on privacy protection, federated learning remains susceptible to poisoning attacks from malicious clients. These attacks can degrade system performance and potentially lead to data privacy breaches. Moreover, real-world IoT datasets are often heterogeneous, further increasing the difficulty of detecting malicious clients. Existing defense mechanisms often struggle to effectively identify malicious clients while maintaining high model performance. To address this issue, we propose a defense mechanism called Malicious client exclusion aggregation (MaEA). This method utilizes KL divergence to preliminarily filter out anomalous clients, aggregates the remaining (preliminarily filtered) clients to obtain a pre-center model, and then identifies and excludes malicious clients by measuring their deviations from this pre-center model. We executed a series of extensive experiments on the CIFAR-10 dataset to demonstrate the effectiveness of MaEA. The results demonstrate that our approach can efficiently detect and identify malicious clients while correcting model performance.

Multimodal Large Language Models (MLLMs) represent the next frontier in artificial intelligence (AI), capable of processing and integrating diverse data types (text, images, audio and video) for richer understanding and generation. However, their potential is limited by privacy constraints: sensitive data resides with different owners who must keep it local, while MLLM parameters themselves require protection. We address this challenge by pioneering complete end-to-end privacy protection that simultaneously secures prompts, multimedia data, and MLLM parameters.In this work, we present PrivMLLM, the first general-purpose secure three-party (3PC) inference framework for MLLMs that serves two data owners while protecting both model parameters and inputs. Our approach is based on three key insights: (1) domain knowledge-aware optimization of fixed-point arithmetic for global performance gains, (2) hybrid secret-sharing protocols that reduce communication and rounds for linear/non-linear operations, and (3) constant-round function secret sharing (FSS)-based protocols for private embedding and maximum.We formally prove PrivMLLM’s security under the rigorous Universal Composability (UC) framework and demonstrate the first end-to-end secure inference system for MLLMs. Experimental results show that our solution enables secure inference for Llama3.2-11B-vision in under 0:5-minute per token, achieving 16 and 8 speedups compared to our implementations built with the SOTA 2PC (CrypTen+) and 3PC (ABY3+) privacypreserving machine learning (PPML) frameworks respectively. Moreover, we benchmark the submodules ViT and LLM against the SOTA schemes, SHAFT (NDSS 2025) and SIGMA (PETS 2024), achieving 1:3~3× performance gains.

Camera-based Semantic Scene Completion (SSC) is to infer the full geometry of objects and scenes from only 2D images. The task is particularly challenging for those in-visible areas, due to the inherent occlusions and lighting ambiguity. Existing works ignore the information missing or ambiguous in those shaded and occluded areas, resulting in distorted geometric prediction. To address this issue, we propose a novel method, Bi-SSC, bidirectional geomet-ric semantic fusion for camera-based 3D semantic scene completion. The key insight is to use the neighboring structure of objects in the image and the spatial differences from different perspectives to compensate for the lack of information in occluded areas. Specifically, we introduce a spatial sensory fusion module with multiple association attention to improve semantic correlation in geometric distributions. This module works within single view and across stereo views to achieve global spatial consistency. Experimental results demonstrate that Bi-SSC outperforms state-of-the-art camera-based methods on SemanticKITTI, particularly excelling in those invisible and shaded areas.

Traditional searchable symmetric encryption (SSE) schemes rarely support context-aware semantic extension, and then lead to the searched results being incomplete or deviating from the user’s query intention. To address this problem, a new context-aware semantically extensible searchable symmetric encryption based on Word2vec model (CASE-SSE) is proposed to achieve context-aware semantic extension in this article. The proposed scheme utilizes outsourced datasets as corpora to extract all keywords for training the Word2vec model, and the trained results is the ontology knowledge base that can be used to extend the semantics of query keywords directly. Further, to facilitate multi-keyword search using the extended query vector, we use the$k$-means clustering algorithm to classify outsourced datasets. We then construct an AVL-tree index and an inverted index based on the classified results, thereby achieving efficient context-aware semantically extensible SSE. The security analysis indicates it is secure and effective. The experimental results show that our scheme is superior in both efficiency and accuracy.