arXiv Science⌕ Search

arXiv subjects

Mamoru Mimura

Publications and source records attributed to Mamoru Mimura.

13 recordsLinked to original sources
arXiv2022

Compiler Provenance Recovery for Multi-CPU Architectures Using a Centrifuge Mechanism

Bit-stream recognition (BSR) has many applications, such as forensic investigations, detection of copyright infringement, and malware analysis. We propose the first BSR that takes a bare input bit-stream and outputs a class label without any preprocessing. To achieve our goal, we propose a centrifuge mechanism, where the upstream layers (sub-net) capture global features and tell the downstream layers (main-net) to switch the focus, even if a part of the input bit-stream has the same value. We applied the centrifuge mechanism to compiler provenance recovery, a type of BSR, and achieved excellent classification. Additionally, downstream transfer learning (DTL), one of the learning methods we propose for the centrifuge mechanism, pre-trains the main-net using the sub-net's ground truth instead of the sub-net's output. We found that sub-predictions made by DTL tend to be highly accurate when the sub-label classification contributes to the essence of the main prediction.

cs.LG↗
arXiv2018

o-glasses: Visualizing x86 Code from Binary Using a 1d-CNN

Malicious document files used in targeted attacks often contain a small program called shellcode. It is often hard to prepare a runnable environment for dynamic analysis of these document files because they exploit specific vulnerabilities. In these cases, it is necessary to identify the position of the shellcode in each document file to analyze it. If the exploit code uses executable scripts such as JavaScript and Flash, it is not so hard to locate the shellcode. On the other hand, it is sometimes almost impossible to locate the shellcode when it does not contain any JavaScript or Flash but consists of native x86 code only. Binary fragment classification is often applied to visualize the location of regions of interest, and shellcode must contain at least a small fragment of x86 native code even if most of it is obfuscated, such as, a decoder for the obfuscated body of the shellcode. In this paper, we propose a novel method, o-glasses, to visualize the shellcode by recognizing the x86 native code using a specially designed one-dimensional convolutional neural network (1d-CNN). The fragment size needs to be as small as the minimum size of the x86 native code in the whole shellcode. Our results show that a 16-instruction-sequence (approximately 48 bytes on average) is sufficient for the code fragment visualization. Our method, o-glasses (1d-CNN), outperforms other methods in that it recognizes x86 native code with a surprisingly high F-measure rate (about 99.95%).

cs.CR↗
arXiv2012

Weyl group invariants

For any odd prime $p$, we prove that the induced homomorphism from the mod $p$ cohomology of the classifying space of a compact simply-connected simple connected Lie group to the Weyl group invariants of the mod $p$ cohomology of the classifying space of its maximal torus is an epimorphism except for the case $p=3$, $G=E_8$.

math.AT↗
arXiv2009

Mui invariants and Milnor operations

We describe Mui invariants in terms of Milnor operations and give a simple proof for Mui's theorem on rings of invariants of polynomial tensor exterior algebras with respect to the action of finite general linear groups. Moreover, we compute some rings of invariants of Weyl groups of maximal non-toral elementary abelian p-subgroups of exceptional Lie groups.

math.AT↗
arXiv2004

Lusternik-Schnirelmann categories of non-simply connected compact simple Lie groups

Let $F \hookrightarrow X \to B$ be a fibre bundle with structure group $G$, where $B$ is $(d{-}1)$-connected and of finite dimension, $d \geq 1$. We prove that the strong L-S category of $X$ is less than or equal to $m + \frac{\dim B}{d}$, if $F$ has a cone decomposition of length $m$ under a compatibility condition with the action of $G$ on $F$. This gives a consistent prospect to determine the L-S category of non-simply connected Lie groups. For example, we obtain $\cat{PU(n)} \leq 3(n{-}1)$ for all $n \geq 1$, which might be best possible, since we have $\cat{\mathrm{PU}(p^r)}=3(p^r{-}1)$ for any prime $p$ and $r \geq 1$. Similarly, we obtain the L-S category of $\mathrm{SO}(n)$ for $n \leq 9$ and $\mathrm{PO}(8)$. We remark that all the above Lie groups satisfy the Ganea conjecture on L-S category.

math.AT↗
arXiv2002

L-S categories of simply-connected compact simple Lie groups of low rank

We determine the L-S category of Sp(3) by showing that the 5-fold reduced diagonal $\widebarΔ_5$ is given by $ν^2$, using a Toda bracket and a generalised cohomology theory $h^*$ given by $h^*(X,A) = \{X/A,{\mathbb S}[0,2]\}$, where ${\mathbb S}[0,2]$ is the 3-stage Postnikov piece of the sphere spectrum ${\mathbb S}$. This method also yields a general result that $\cat(Sp(n)) \geq n+2$ for $n \geq 3$, which improves the result of Singhof.

math.AT↗