Arrow Research search

Author name cluster

Mohammad Akbari

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

5 papers
1 author row

Possible papers

5

AAAI Conference 2025 Conference Paper

Task-Agnostic Language Model Watermarking via High Entropy Passthrough Layers

  • Vaden Masrani
  • Mohammad Akbari
  • David Ming Xuan Yue
  • Ahmad Rezaei
  • Yong Zhang

In the era of costly pre-training of large language models, ensuring the intellectual property rights of model owners, and insuring that said models are responsibly deployed, is becoming increasingly important. To this end, we propose model watermarking via passthrough layers, which are added to existing pre-trained networks and trained using a self-supervised loss such that the model produces high-entropy output when prompted with a unique private key, and acts normally otherwise. Unlike existing model watermarking methods, our method is fully task-agnostic, and can be applied to both classification and sequence-to-sequence tasks without requiring advanced access to downstream fine-tuning datasets. We evaluate the proposed passthrough layers on a wide range of downstream tasks, and show experimentally our watermarking method achieves a near-perfect watermark extraction accuracy and false-positive rate in most cases without damaging original model performance. Additionally, we show our method is robust to both downstream fine-tuning, fine-pruning, and layer removal attacks, and can be trained in a fraction of the time required to train the original model. Code is available in the paper.

PDF Details DOI

AAAI Conference 2021 Conference Paper

Learned Bi-Resolution Image Coding using Generalized Octave Convolutions

  • Mohammad Akbari
  • Jie Liang
  • Jingning Han
  • Chengjie Tu

Learned image compression has recently shown the potential to outperform the standard codecs. State-of-the-art ratedistortion (R-D) performance has been achieved by contextadaptive entropy coding approaches in which hyperprior and autoregressive models are jointly utilized to effectively capture the spatial dependencies in the latent representations. However, the latents are feature maps of the same spatial resolution in previous works, which contain some redundancies that affect the R-D performance. In this paper, we propose a learned bi-resolution image coding approach that is based on the recently developed octave convolutions to factorize the latents into high and low resolution components. Therefore, the spatial redundancy is reduced, which improves the R-D performance. Novel generalized octave convolution and octave transposed-convolution architectures with internal activation layers are also proposed to preserve more spatial structure of the information. Experimental results show that the proposed scheme outperforms all existing learned methods as well as standard codecs such as the next-generation video coding standard VVC (4: 2: 0) in both PSNR and MS-SSIM. We also show that the proposed generalized octave convolution can improve the performance of other auto-encoder-based schemes such as semantic segmentation and image denoising.

PDF Details

AAAI Conference 2018 System Paper

Vertical Domain Text Classification: Towards Understanding IT Tickets Using Deep Neural Networks

  • Jianglei Han
  • Mohammad Akbari

It is challenging to directly apply text classification models without much feature engineering on domain-specific use cases, and expect the state of art performance. Much more so when the number of classes is large. Convolutional Neural Network (CNN or ConvNet) has attracted much in text mining due to its effectiveness in automatic feature extraction from text. In this paper, we compare traditional and deep learning approaches for automatic categorization of IT tickets in a real world production ticketing system. Experimental results demonstrate the good potential of CNN models in our task.

PDF Details