Published in SIGIR 2026, 2026
A SIGIR 2026 full paper on continual generative retrieval with parametric memory.
Recommended citation: Mekonnen, K. A., et al. (2026). "A Parametric Memory Head for Continual Generative Retrieval." Proceedings of SIGIR 2026.
Download Paper
Published in SIGIR 2026 Reproducibility Track, 2026
A SIGIR 2026 Reproducibility Track paper stress-testing the look-ahead prior in generative retrieval.
Recommended citation: Mekonnen, K. A., et al. (2026). "Lost in Decoding? Reproducing and Stress-Testing the Look-Ahead Prior in Generative Retrieval." Proceedings of SIGIR 2026.
Download Paper
Published in ACL Findings 2025, 2025
An ACL Findings 2025 paper on embeddings and benchmarks for Amharic passage retrieval.
Recommended citation: Mekonnen, K. A., et al. (2025). "Optimized Text Embeddings & Benchmarks for Amharic Passage Retrieval." Findings of the Association for Computational Linguistics.
Download Paper
Published in SIGIR 2025, 2025
A SIGIR 2025 full paper on lightweight document relevance optimization for generative information retrieval.
Recommended citation: Mekonnen, K. A., et al. (2025). "Lightweight and Direct Document Relevance Optimization for Generative Information Retrieval." Proceedings of SIGIR 2025.
Download Paper
Published in arXiv, 2024
An arXiv preprint on conditioning GANs without a training dataset.
Recommended citation: Mekonnen, K. A., et al. (2024). "Conditioning GAN Without Training Dataset." arXiv preprint.
Download Paper
Published in arXiv, 2024
An arXiv preprint on adversarial knowledge distillation for faster diffusion sampling.
Recommended citation: Mekonnen, K. A., et al. (2024). "Adv-KD: Adversarial Knowledge Distillation for Faster Diffusion Sampling." arXiv preprint.
Download Paper
Published in arXiv, 2023
An arXiv preprint on guiding StyleGAN to create labeled synthetic face datasets for underrepresented groups.
Recommended citation: Mekonnen, K. A., et al. (2023). "Balanced Face Dataset: Guiding StyleGAN for Labeled Synthetic Face Image Dataset for Underrepresented Group." arXiv preprint.
Download Paper