Searcharxiv⌕ Search

arXiv subjects

Satoshi Matsuura

Publications and source records attributed to Satoshi Matsuura.

4 recordsLinked to original sources

From Token Changes to Task Failures: Bit Flips in a Shared KV-Cache Block

In LLM serving systems with prefix caching, each KV block of a shared system prompt is held in GPU memory as a single physical copy, which every request with that prompt reads. Our earlier work measured change and showed that one bit flip in such a block changes many later responses without making them visibly degenerate. This paper extends that work by measuring how often the changed responses are wrong. Using software fault injection into vLLM serving Qwen3-8B, we flip the most significant mantissa bit or the sign bit of a value in such a block. We score responses on four tasks: code generation, short-answer questions, multiple-choice reasoning, and refusal under a policy. The fraction of changes that turned a correct response wrong depended strongly on the task: 0.023 on MBPP+ and 0.59 on held-out TriviaQA at the mantissa bit, and 0.019 and 0.60 at the sign bit. Harm fell on requests whose sampled answers are sometimes wrong, so the damage of a flip depends on how common such requests are. Harmed responses looked like ordinary errors, so a flip should be detected by checking cached blocks rather than by monitoring outputs.

cs.CR↗

Detecting Generated-Depth Injection Attacks on RGB-D Place Authentication

Advances in monocular depth estimation (MDE) now let an attacker generate depth from publicly available RGB images and submit forged RGB-D queries. Such attacks break a core assumption of RGB-D place authentication: that depth comes from a sensor. We study a presentation attack detection (PAD) classifier that uses geometric consistency between the enrollment and query views, as a software-layer defense. A three-stage evaluation shows how far this classifier detects software-injected depth, depending on how accurately the attacker knows the camera intrinsics. In the first stage, the attacker has only an estimate of the intrinsics, and the PAD classifier separates forgeries generated by all three MDE models with a validation area under the receiver operating characteristic curve (AUC) of 0.997-0.999. In the second stage, the attacker obtains the exact intrinsics, and the original PAD classifier falls to near chance (AUC 0.54-0.61); a retrained PAD classifier, however, still has the registration residual between enrollment and query as a cue. In the third stage, an oracle evaluation, in which the evaluator fits warps using the ground-truth 3D geometry, shows that detection falls to near chance only in its strongest case, where every valid query keypoint is given its ground-truth 3D coordinate. Of the detection cues, the reprojection error captures the attacker's misestimate of the intrinsics, whereas the registration residual remains even when the attacker knows the exact intrinsics.

cs.CR↗

Bit-Flip Vulnerability of Shared KV-Cache Blocks in LLM Serving Systems

Rowhammer on GPU DRAM has enabled adversarial bit flips in model weights; shared KV-cache blocks in LLM serving systems present an analogous but previously unexamined target. In vLLM's Prefix Caching, these blocks exist as a single physical copy without integrity protection. Using software fault injection under ideal bit targeting, we characterize worst-case severity and identify three properties: (1) Silent divergence - 13 of 16 BF16 bit positions produce coherent but altered outputs, indistinguishable from legitimate responses without a clean baseline. (2) Selective propagation - only requests sharing the targeted prefix are affected. (3) Persistent accumulation - no temporal decay occurs, so cumulative damage grows linearly with subsequent requests. Together, these constitute a threat profile distinct from weight corruption: silent divergence and selective propagation enable detection evasion; persistent accumulation then proceeds unchecked, yielding damage amplification bounded only by how long the block remains cached. A checksum-based countermeasure detects any single-bit corruption at scheduling time, bounding cumulative damage to one batch independent of the block's cache lifetime, with negligible overhead. These results argue for integrity protection of prefix blocks before end-to-end exploitation is demonstrated.

cs.CR↗

On the Experimental Evaluation of Vehicular Networks: Issues, Requirements and Methodology Applied to a Real Use Case

One of the most challenging fields in vehicular communications has been the experimental assessment of protocols and novel technologies. Researchers usually tend to simulate vehicular scenarios and/or partially validate new contributions in the area by using constrained testbeds and carrying out minor tests. In this line, the present work reviews the issues that pioneers in the area of vehicular communications and, in general, in telematics, have to deal with if they want to perform a good evaluation campaign by real testing. The key needs for a good experimental evaluation is the use of proper software tools for gathering testing data, post-processing and generating relevant figures of merit and, finally, properly showing the most important results. For this reason, a key contribution of this paper is the presentation of an evaluation environment called AnaVANET, which covers the previous needs. By using this tool and presenting a reference case of study, a generic testing methodology is described and applied. This way, the usage of the IPv6 protocol over a vehicle-to-vehicle routing protocol, and supporting IETF-based network mobility, is tested at the same time the main features of the AnaVANET system are presented. This work contributes in laying the foundations for a proper experimental evaluation of vehicular networks and will be useful for many researchers in the area.

cs.NI↗