Recent advances in Large Language Models (LLMs) have yielded impressive successes on many language tasks. However, efficient processing of long contexts using LLMs remains a significant challenge. We introduce EpMAN – a method for processing long contexts in an episodic memory module while holistically attending to semantically-relevant context chunks. Output from episodic attention is then used to reweigh the decoder’s self-attention to the stored KV cache of the context during training and generation. When an LLM decoder is trained using EpMAN, its performance on multiple challenging single-hop long-context recall and question-answering benchmarks is found to be stronger and more robust across the range from 16k to 256k tokens than baseline decoders trained with self-attention, and popular retrieval-augmented generation frameworks.

Efficient and accurate updating of knowledge stored in Large Language Models (LLMs) is one of the most pressing research challenges today. This paper presents Larimar - a novel, brain-inspired architecture for enhancing LLMs with a distributed episodic memory. Larimar’s memory allows for dynamic, one-shot updates of knowledge without the need for computationally expensive re-training or fine-tuning. Experimental results on multiple fact editing benchmarks demonstrate that Larimar attains accuracy comparable to most competitive baselines, even in the challenging sequential editing setup, but also excels in speed—yielding speed-ups of 8-10x depending on the base LLM —as well as flexibility due to the proposed architecture being simple, LLM-agnostic, and hence general. We further provide mechanisms for selective fact forgetting, information leakage prevention, and input context length generalization with Larimar and show their effectiveness. Our code is available at https://github.com/IBM/larimar.

Subhajit Chaudhury, Sarathkrishna Swaminathan, Daiki Kimura, Prithviraj Sen, Keerthiram Murugesan, Rosario Uceda-Sosa, Michiaki Tatsubori, Achille Fokoue, Pavan Kapanipathi, Asim Munawar, Alexander Gray. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

Subhajit Chaudhury, Sarathkrishna Swaminathan, Chulaka Gunasekara, Maxwell Crouse, Srinivas Ravishankar, Daiki Kimura, Keerthiram Murugesan, Ramón Fernandez Astudillo, Tahira Naseem, Pavan Kapanipathi, Alexander Gray. Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing. 2022.

An advertiser usually provides a specific demographic target to a content publisher. But due to the limited profile information on each viewer, the publisher suffers from low accuracy in targeting. Once an advertisement is shown to viewers, a third party (like Nielsen, Comscore) validates how close the publisher was to the target audience. Publisher also receives the demographic mix of each show from the third party. Low accuracy in targeting is expensive as the publisher gets paid only for the impressions which were on target. In this work, we propose a new approach to incorporate user level latent features developed from the show-wise demographic mix provided by the third party, along with session level features of viewers to improve the demographic predictions. In congruence with current industry practice, we train our model on a small labeled demographic segment and establish the effectiveness of our approach over existing approaches through experiments.

Multiple Kernel Learning, or MKL, extends (kernelized) SVM by attempting to learn not only a classifier/regressor but also the best kernel for the training task, usually from a combination of existing kernel functions. Most MKL methods seek the combined kernel that performs best over every training example, sacrificing performance in some areas to seek a global optimum. Localized kernel learning (LKL) overcomes this limitation by allowing the training algorithm to match a component kernel to the examples that can exploit it best. Several approaches to the localized kernel learning problem have been explored in the last several years. We unify many of these approaches under one simple system and design a new algorithm with improved performance. We also develop enhanced versions of existing algorithms, with an eye on scalability and performance.