The ability to base current computations on memories from the past is critical for many cognitive tasks such as story understanding.

The ability to base current computations on memories from the past is critical for many cognitive tasks such as story understanding. Hebbian-type synaptic plasticity is believed to underlie the retention of memories over medium and long time scales in the brain. However, it is unclear how such plasticity processes are integrated with computations in cortical networks. Here, we propose Hebbian Memory Networks (H-Mems), a simple neural network model that is built around a core hetero-associative network subject to Hebbian plasticity. We show that the network can be optimized to utilize the Hebbian plasticity processes for its computations. H-Mems can one-shot memorize associations between stimulus pairs and use these associations for decisions later on. Furthermore, they can solve demanding question-answering tasks on synthetic stories. Our study shows that neural network models are able to enrich their computations with memories through simple Hebbian plasticity processes.

Here we give details to our models, and to the encoding and representation of images and questions used in our models. We used a CNN as input encoder in this task. The last fully connected layer was of size 128 followed by a ReLU nonlinearity (BatchNorm denotes a batch normalization layer [2]). Training examples were generated as described in the main text. Here, an example is one full sequence of image pairs (including random images) and one query image.

Here we give details to our models, and to the encoding and representation of images and questions used in our models. We used a CNN as input encoder in this task. The last fully connected layer was of size 128 followed by a ReLU nonlinearity (BatchNorm denotes a batch normalization layer [2]). Training examples were generated as described in the main text. Here, an example is one full sequence of image pairs (including random images) and one query image.

Long-term memory is one of the key factors influencing the reasoning capabilities of Large Language Model Agents (LLM Agents). Incorporating a memory mechanism that effectively integrates past interactions can significantly enhance decision-making and contextual coherence of LLM Agents. While recent works have made progress in memory storage and retrieval, such as encoding memory into dense vectors for similarity-based search or organizing knowledge in the form of graph, these approaches often fall short in structured memory organization and efficient retrieval. To address these limitations, we propose a Hierarchical Memory (H-MEM) architecture for LLM Agents that organizes and updates memory in a multi-level fashion based on the degree of semantic abstraction. Each memory vector is embedded with a positional index encoding pointing to its semantically related sub-memories in the next layer. During the reasoning phase, an index-based routing mechanism enables efficient, layer-by-layer retrieval without performing exhaustive similarity computations. We evaluate our method on five task settings from the LoCoMo dataset. Experimental results show that our approach consistently outperforms five baseline methods, demonstrating its effectiveness in long-term dialogue scenarios.

The ability to base current computations on memories from the past is critical for many cognitive tasks such as story understanding. Hebbian-type synaptic plasticity is believed to underlie the retention of memories over medium and long time scales in the brain. However, it is unclear how such plasticity processes are integrated with computations in cortical networks. Here, we propose Hebbian Memory Networks (H-Mems), a simple neural network model that is built around a core hetero-associative network subject to Hebbian plasticity. We show that the network can be optimized to utilize the Hebbian plasticity processes for its computations.

The motivation of the model is unclear. In other words, why can this model work on the two tasks? We cannot simply say it uses Hebbian rule which agrees with biological system then it should work. A reason, or intuition, from the perspective of machine learning should be provided. I want to see explanations on both tasks in the rebuttal.

The authors propose a novel architecture to solve memory-based question and answering task. The key idea is to use a hetero-associative memory that utilizes fast Hebbian learning. All the reviewers agree that this paper makes a strong contribution to the literature and the experimental results are sufficient to support the claims made in the paper. Based on the clarification questions sought by the reviewers, the writing could use improvement and authors have promised to address the issues raised in the additional space they will have in the final submission.

The ability to base current computations on memories from the past is critical for many cognitive tasks such as story understanding. Hebbian-type synaptic plasticity is believed to underlie the retention of memories over medium and long time scales in the brain. However, it is unclear how such plasticity processes are integrated with computations in cortical networks. Here, we propose Hebbian Memory Networks (H-Mems), a simple neural network model that is built around a core hetero-associative network subject to Hebbian plasticity. We show that the network can be optimized to utilize the Hebbian plasticity processes for its computations.