Large Multimodal Models (LMMs) extend Large Language Models (LLMs) by handling diverse inputs such as images, audio, and video, but at the cost of adding a multimodal encoding stage that increases both computational and memory overhead. This step negatively affects key Service Level Objectives (SLOs), such as time to first token (TTFT) and time per output token (TPOT). We introduce Encode-Prefill-Decode (EPD) Disaggregation, a novel framework that separates the encoding, prefill, and decode stages onto dedicated resources. Unlike current systems, which bundle encoding and prefill together, our approach decouples these steps, unlocking new opportunities and optimizations. These include a mechanism to cache multimedia tokens for efficient transfer, a novel way to parallelize the encoding load within a request, a module for optimal resource allocation for disaggregated serving, and a novel role-switching method to handle changing workload characteristics. Experimental evaluations with popular LMMs show substantial gains in memory efficiency (up to 15$\times$ lower peak memory utilization), batch sizes (up to 22$\times$ larger), 10$\times$ more images per request, and 2.2$\times$ larger KV caches. Furthermore, it leads to significant improvements in SLO attainment (up to 90–100% improvement) and TTFT (up to 71% reduction), compared to systems that do not disaggregate. The code is available at https://github.com/vbdi/epdserve.

Symbol recovery in reverse engineering is crucial for restoring variable and data structure information in compiled binaries. While learning-based methods have shown promise in recovering both semantic information (names and types) and syntactic information (shapes), they require comprehensive datasets where expressions in binary code are precisely aligned with their source code equivalents. Current techniques for generating such alignments struggle with complex data access patterns, resulting in incomplete training data and consequently hampering model performance and recovery accuracy. We present AST-Align, a novel technique unifying alignment of variables and struct access expressions across multiple architectures (x86 and ARM) and languages (C/C++/Rust). AST-Align significantly improves the number of generated ground truths, capturing four times more struct fields than previous methods. Using this algorithm, we develop R3-Bench, a metadata-rich, extensible dataset with explicit project inclusion criteria and reproducible processing pipeline, comprising over 10 million functions across multiple architectures. Our evaluation establishes baseline performance by testing various approaches from n-gram models to Large Language Models. The results show that while general LLMs initially perform poorly, their effectiveness dramatically improves with proper demonstration. R3-Bench provides a robust foundation for assessing model capabilities and serves as a valuable reference for future symbol recovery research.

This paper presents a Simple and effective unsupervised adaptation method for Robust Object Detection (SimROD). To overcome the challenging issues of domain shift and pseudo-label noise, our method integrates a novel domain-centric data augmentation, a gradual self-labeling adaptation procedure, and a teacher-guided fine-tuning mechanism. Using our method, target domain samples can be leveraged to adapt object detection models without changing the model architecture or generating synthetic data. When applied to image corruptions and high-level cross-domain adaptation benchmarks, our method outperforms prior baselines on multiple domain adaptation benchmarks. SimROD achieves new state-of-the-art on standard real-to-synthetic and cross-camera setup benchmarks. On the image corruption benchmark, models adapted with our method achieved a relative robustness improvement of 15-25% AP50 on Pascal-C and 5-6% AP on COCO-C and Cityscapes-C. On the cross-domain benchmark, our method outperformed the best baseline performance by up to 8% and 4% AP50 on Comic and Watercolor respectively.1

Drivers on Apple OSes (e.g., iOS, tvOS, iPadOS, macOS, etc.) run in the kernel space and driver vulnerabilities can incur serious security consequences. A recent report from Google Project Zero shows that driver vulnerabilities on Apple OSes have been actively exploited in the wild. Also, we observed that driver vulnerabilities have accounted for one-third of kernel bugs in recent iOS versions based on Apple's security updates. Despite the serious security implications, systematic static analysis on Apple drivers for finding security vulnerabilities has never been done before, not to mention any large-scale study of Apple drivers.

The use of uninitialized variables is a common issue. It could cause kernel information leak, which defeats the widely deployed security defense, i.e., kernel address space layout randomization (KASLR). Though a recent system called Bochspwn Reloaded reported multiple memory leaks in Windows kernels, how to effectively detect this issue is still largely behind.

As a critical feature for enhancing user experience, cross-app URL invocation has been reported to cause unauthorized execution of app components. Although protection has already been put in place, little has been done to understand the security risks of navigating an app's WebView through an URL, a legitimate need for displaying the app's UI during cross-app interactions. In our research, we found that the current design of such cross-WebView navigation actually opens the door to a cross-app remote infection, allowing a remote adversary to spread malicious web content across different apps' WebView instances and acquire stealthy and persistent control of these apps. This new threat, dubbed Cross-App WebView Infection (XAWI), enables a series of multi-app, colluding attacks never thought before, with significant real world impacts. Particularly, we found that the remote adversary can collectively utilize multiple infected apps' individual capabilities to escalate his privileges on a mobile device or orchestrate a highly realistic remote Phishing attack (e.g., running a malicious script in Chrome to stealthily change Twitter's WebView to fake Twitter's own login UI). We show that the adversary can easily find such attack "building blocks" (popular apps whose WebViews can be redirected by another app) through an automatic fuzz, and discovered about 7.4% of the most popular apps subject to the XAWI attacks, including Facebook, Twitter, Amazon and others. Our study reveals the contention between the demand for convenient cross-WebView communication and the need for security control on the channel, and makes the first step toward building OS-level protection to safeguard this fast-growing technology.

Privacy inference attacks based on sensor data is an emerging and severe threat on smart devices, in which malicious applications leverage data from innocuous sensors to infer sensitive information of user, e.g., utilizing accelerometers to infer user’s keystroke. In this paper, we present Sensor Guardian, a privacy protection system that mitigates this threat on Android by hooking and controlling applications’ access to sensors. Sensor Guardian inserts hooks into applications by statically instrumenting their APK (short for Android Package Kit) files and enforces control policies in these hooks at runtime. Our evaluation shows that Sensor Guardian can effectively and efficiently mitigate the privacy inference threat on Android sensors, with negligible overhead during both static instrumentation and runtime control.

With the popularity of today's usability-oriented designs, dubbed Zero Configuration or ZeroConf, unclear are the security implications of these automatic service discovery, "plug-and-play" techniques. In this paper, we report the first systematic study on this issue, focusing on the security features of the systems related to Apple, the major proponent of ZeroConf techniques. Our research brings to light a disturbing lack of security consideration in these systems' designs: major ZeroConf frameworks on the Apple platforms, including the Core Bluetooth Framework, Multipeer Connectivity and Bonjour, are mostly unprotected and popular apps and system services, such as Tencent QQ, Apple Handoff, printer discovery and AirDrop, turn out to be completely vulnerable to an impersonation or Man-in-the-Middle (MitM) attack, even though attempts have been made to protect them against such threats. The consequences are serious, allowing a malicious device to steal the user's SMS messages, email notifications, documents to be printed out or transferred to another device. Most importantly, our study highlights the fundamental security challenges underlying ZeroConf techniques: in the absence of any pre-configured secret across different devices, authentication has to rely on Apple-issued public-key certificate, which however cannot be properly verified due to the difficulty in finding a unique, nonsensitive and widely known identity of a human user to bind her to her certificate. To address this issue, we developed a suite of new techniques, including a conflict detection approach and a biometric technique that enables the user to speak out her certificate through 6 distinct, rare but pronounceable words to let those who know her voice verify her certificate. We performed a security analysis on the new protection and evaluated its usability and effectiveness using two user studies involving 60 participants. Our research shows that the new protection fits well with the existing ZeroConf systems such as AirDrop. It is well received by users and also providing effective defense even against recently proposed speech synthesis attacks.

On modern operating systems, applications under the same user are separated from each other, for the purpose of protecting them against malware and compromised programs. Given the complexity of today's OSes, less clear is whether such isolation is effective against different kind of cross-app resource access attacks (called XARA in our research). To better understand the problem, on the less-studied Apple platforms, we conducted a systematic security analysis on MAC OS~X and iOS. Our research leads to the discovery of a series of high-impact security weaknesses, which enable a sandboxed malicious app, approved by the Apple Stores, to gain unauthorized access to other apps' sensitive data. More specifically, we found that the inter-app interaction services, including the keychain, WebSocket and NSConnection on OS~X and URL Scheme on the MAC OS and iOS, can all be exploited by the malware to steal such confidential information as the passwords for iCloud, email and bank, and the secret token of Evernote. Further, the design of the app sandbox on OS~X was found to be vulnerable, exposing an app's private directory to the sandboxed malware that hijacks its Apple Bundle ID. As a result, sensitive user data, like the notes and user contacts under Evernote and photos under WeChat, have all been disclosed. Fundamentally, these problems are caused by the lack of app-to-app and app-to-OS authentications. To better understand their impacts, we developed a scanner that automatically analyzes the binaries of MAC OS and iOS apps to determine whether proper protection is missing in their code. Running it on hundreds of binaries, we confirmed the pervasiveness of the weaknesses among high-impact Apple apps. Since the issues may not be easily fixed, we built a simple program that detects exploit attempts on OS~X, helping protect vulnerable apps before the problems can be fully addressed.

For computer vision-based inspection of electronic chips or dies in semiconductor production lines, we propose a new method to effectively and efficiently detect defects in images. Different from the traditional methods that compare the image of each test chip or die with the template image one by one, which are sensitive to misalignment between the test and template images, a collection of multiple test images are used as the input image for processing simultaneously in our method with two steps. The first step is to obtain salient regions of the whole collection of test images, and the second step is to evaluate local discrepancy between salient regions in test images and the corresponding regions in the defect-free template image. To be more specific, in the first step of our method, phase-only Fourier transform (POFT), which is computationally efficient for online applications in industry, is used for saliency detection. We provide the theoretical justification for POFT to be effective to attenuate the normal regions and amplify the defects in multiple test images, which are usually arranged in a matrix format in industrial practice. By comparing with four other popular methods, the proposed algorithm can efficiently accommodate small variations (inevitable in practice) in test chips or dies, such as the spatial misalignments and product variations. Experimental results on a large-scale database including 1073 images, 94 of which are defective, show that our method performs much better than the other methods in terms of precision, recall, and F-measure.