Byte-level generative predictions for forensics multimedia carving

Jaewon Lee; Md Eimran Hossain Eimon; Avinash Srinivasan; Hari Kalva

cs.CVApr 13, 2026

Byte-level generative predictions for forensics multimedia carving

Authors: Jaewon Lee, Md Eimran Hossain Eimon, Avinash Srinivasan, Hari Kalva

Organizations: United States Naval Academy, 121 Blake Road, Annapolis, USA · Florida Atlantic University, 777 Glades Road, Boca Raton, USA

Abstract

Digital forensic investigations often face significant challenges when recovering fragmented multimedia files that lack file system metadata. While traditional file carving relies on signatures and discriminative deep learning models for fragment classification, these methods cannot reconstruct or predict missing data. We propose a generative approach to multimedia carving using bGPT, a byte-level transformer designed for next-byte prediction. By feeding partial BMP image data into the model, we simulate the generation of likely fragment continuations. We evaluate the fidelity of these predictions using different metrics, namely, cosine similarity, structural similarity index (SSIM), chi-square distance, and Jensen-Shannon divergence (JSD). Our findings demonstrate that generative models can effectively predict byte-level patterns to support fragment matching in unallocated disk space.

Figures & tables

Figure 1: The bGPT framework simulates digital systems through native binary data and integrates diverse data types into a single model, treating everything as a byte sequence. [ 10 ] .

Figure 2: Flow diagram of a pseudo-simulation pipeline where fragmented BMP inputs are processed by bGPT to predict the next fragment, which is compared to a mixed multimedia fragment pool using similarity metrics to identify the top five matches.

Metric	Mean			Std Dev			95% CI (±)
Metric	$P_{1}$ ( $\sim$ 1.8 kB)	$P_{2}$ ( $\sim$ 1.2 kB)	$P_{3}$ ( $\sim$ 0.6 kB)	$P_{1}$ ( $\sim$ 1.8 kB)	$P_{2}$ ( $\sim$ 1.2 kB)	$P_{3}$ ( $\sim$ 0.6 kB)	$P_{1}$ ( $\sim$ 1.8 kB)	$P_{2}$ ( $\sim$ 1.2 kB)	$P_{3}$ ( $\sim$ 0.6 kB)
Chi-square	409.58	335.61	174.97	161.23	139.24	61.96	23.12	19.96	8.88
Cosine	0.6690	0.6749	0.7641	0.1735	0.1605	0.1198	0.0249	0.0230	0.0172
JSD	0.1864	0.2338	0.2458	0.0798	0.1050	0.0938	0.0114	0.0151	0.0134
SSIM	0.0856	0.1016	0.3262	0.0738	0.0966	0.1628	0.0106	0.0138	0.0233

Table 1: Results of similarity analysis between predicted fragments and their real continuations across fragment sizes.

Figure 3: Box plots showing similarity scores between predicted and real fragments across three predicted fragment sizes: $P_{1}$ (1.8 kB), $P_{2}$ (1.2 kB), and $P_{3}$ (0.6 kB), using four evaluation metrics: cosine similarity, SSIM, chi-square distance, and Jensen–Shannon divergence. Each plot captures the distribution and variation in similarity across different prediction groups.

Figure 4: SSIM heatmap between predicted and real image fragments from Fig. 5 . Local SSIM values were computed using a $7\times 7$ sliding window, where each pixel represents the similarity of its local neighborhood.

Figure 5: Example of fragment reconstruction using bGPT. The input fragment (a) is provided to the model to generate the predicted continuation (b), which approximates the true continuation (c). Combining (a) and (b) produces the reconstructed image (d), compared against the original full image (e).

Ranking Outcome	Count
Correct match ranked 1st	20
Correct match within top 5 (not 1st)	4
Correct match not within top 5	6
Total predictions	30

Table 2: Fragment matching results.

Digital Forensics

Digital forensics develops methods to recover, authenticate, and interpret digital evidence while preserving its integrity for investigations.

See all 71 papers

Dynamic Multi-Byte Prediction With Hierarchical Language Models
Aug 16, 2026Abraham Toluwase Owodunni, Chibuzor Okocha, Christan Grant +2Language Modeling LLM Inference Acceleration
The Ohio State University · University of Florida · University of Washington
Beyond Surface Artifacts: Capturing Shared Latent Forgery Knowledge Across Modalities
Apr 9, 2026Jingtong Dou, Chuancheng Shi, Jian Wang +3Multimodal Deepfake Detection Multimodal Domain Generalization
The University of Sydney · Nanjing University of Posts and Telecommunications · National University of Singapore
LaP-Forensics: Latent-Pixel Consistency Guided Multimodal Reasoning for Deepfake Detection
Jul 28, 2026Can Wang, Yuhao Wang, Yushe Cao +2Multimodal Deepfake Detection Digital Image Forensics
The Hong Kong Polytechnic University, Hong Kong, China · University College London, London, United Kingdom · Tsinghua University, Beijing, China +2

Aug 16, 2026J/K move · Enter open · S save

Dynamic Multi-Byte Prediction With Hierarchical Language Models

Abraham Toluwase Owodunni, Chibuzor Okocha, Christan Grant, Tomasz Limisiewicz+1

The Ohio State University · University of Florida · University of Washington

Language Modeling LLM Inference Acceleration Byte-Level Language Model Multi-Token Prediction

Open paper AskPDF arXiv

Byte-level hierarchical language models (LMs) have recently emerged as a robust alternative to their popular counterparts that use subword tokenization. However, generating one byte at a time remains a bottleneck for inference speed. To address this, we introduce multi-byte prediction (MBP), which generates multiple bytes in parallel, speeding up inference with minimal performance impact and no additional parameters. MBP builds on the popular multi-token prediction (MTP) paradigm with two crucial innovations. First, we introduce a variable-length prediction window that aligns with the latent tokens, or segments, of a hierarchical LM. Second, we implement a novel attention-masking scheme that enables parallel byte prediction without violating causality. We show that multi-byte prediction strikes a Pareto-optimal trade-off across multiple generative tasks, instruction following, question answering, summarization, and machine translation, achieving the best trade-off between performance and inference throughput.

Abstract

Digital Forensics

Explore similar work

Dynamic Multi-Byte Prediction With Hierarchical Language Models

Beyond Surface Artifacts: Capturing Shared Latent Forgery Knowledge Across Modalities

LaP-Forensics: Latent-Pixel Consistency Guided Multimodal Reasoning for Deepfake Detection