24 papers · ranked by Valyu relevance
Chen Fang, Gourav Datta, Mujahid Al Rafi, Hyeran Jeon + 1 more
The expansion of neural network sizes and the enhancement of image resolution through modern camera sensors result in heightened memory and power demands for neural networks. Reducing peak memory, which is the maximum memory consumed during the execution of a neural network, is critical to deploy neural networks on…
Zidi Xiong, Yuping Lin, Wenya Xie, Pengfei He + 3 more
'Himabindu Lakkaraju' 'Zhen Xiang'] Memory is a critical component in large language model (LLM)-based agents, enabling them to store and retrieve past executions to improve task performance over time. In this paper, we conduct an empirical study on how memory management choices impact the LLM agents' behavior…
Mitsuru Kusumoto, Takuya Inoue, Gentaro Watanabe, Takuya Akiba + 1 more
'Masanori Koyama'] Recomputation algorithms collectively refer to a family of methods that aims to reduce the memory consumption of the backpropagation by selectively discarding the intermediate results of the forward propagation and recomputing the discarded results as needed. In this paper, we will propose a novel…
Zeyuan Ye, Haoran Li, Liang Tian, Changsong Zhou
Understanding how the brain preserves information despite intrinsic noise is a fundamental question in working memory. Typical working memory tasks consist of delay phase for maintaining information, and decoding phase for retrieving information. While previous works have focused on the delay neural dynamics, it is…
Md. Shamim Hossain, Shiro Mawatari, Takehiko Fujino
Plasmalogens (Pls), a kind of glycerophospholipids, have shown potent biological effects but their role in hippocampus-dependent memory remained mostly elusive. Here, we first report Pls can enhance endogenous expression of brain-derived neurotrophic factor (Bdnf) in the hippocampus and promotes neurogenesis associated…
Deliang Wen, Ke Sun
Large language models (LLMs) face inherent limitations in memory, including restricted context windows [[1]], long-term knowledge forgetting, redundant information accumulation, and hallucination generation [[2]], all of which severely constrain sustained dialogue and personalized services [[3]]. This paper proposes…
David Shih
Integration-by-parts (IBP) reduction of Feynman integrals to master integrals is a key computational bottleneck in precision calculations in high-energy physics. Traditional approaches based on the Laporta algorithm require solving large systems of equations, leading to memory consumption that grows rapidly with…
Wayne S. Sossin
Synapses are diverse in form and function. While there are strong evidential and theoretical reasons for believing that memories are stored at synapses, the concept of a specialized “memory synapse” is rarely discussed. Here, we review the evidence that memories are stored at the synapse and consider the opposing…
Kwan Tung Li, Daoyun Ji, Changsong Zhou
Neurodegeneration is a characteristic of Alzheimer’s disease (AD), but its effect on neural activity dynamics underlying memory deficits is unclear. Here, we studied the effects of synaptic degeneration on neural activities associated with memory recall, memory rescue by slow-gamma stimulation, and learning a new…
Santiago Abel Merlo, Mariano Andrés Belluscio, Maria Eugenia Pedreira, Emiliano Merlo
'Emiliano Merlo'] Memory persistence is a double edge sword. Persistence of adaptive memories is essential for survival and even determines who we are. Neurodegenerative conditions with significant memory loss such as Alzheimer’s disease, testify how defects of memory persistence have severe and irreversible effects on…
Junjiro Horiuchi, Nozomi Uemura, Shiro Horiuchi, Minoru Saitoe
While forgetting has been studied extensively in various organisms, its precise nature has often been unclear. Here, we used behavioral experiments in Drosophila to determine that a significant aspect of forgetting consists of a decrease in the ability of a memory to induce an appropriate behavior. We tested flies for…
Ahmad Beirami, Mohsen Sardari, Faramarz Fekri
—The existence of considerable amount of redundancy in the Internet traffic at the packet level has stimulated the deployment of packet-level redundancy elimination techniques within the network by enabling network nodes to memorize data packets. Redundancy elimination results in traffic reduction which in turn…
Jacob A. Berry, Dana C. Guhle, Ronald L. Davis
Recent and pioneering animal research has revealed the brain utilizes a variety of molecular, cellular, and network-level mechanisms used to forget memories in a process referred to as “active forgetting”. Active forgetting increases behavioral flexibility and removes irrelevant information. Individuals with impaired…
Michał J. Wójcik, Amy Li, Dante Wasmuht, Jake P. Stroud + 3 more
Working memory has been traditionally studied as a passive storage for information. However, recent advances have suggested that working memory is prospective rather than retrospective, meaning that its content undergoes transformations that will support future behaviour. One perspective that underscores this notion…
Authors not listed
Accurate prediction of chemical reaction yields remains essential for accelerating synthesis optimization, yet current machine learning models face critical limitations in capturing temporal dynamics, providing calibrated uncertainty estimates, and explicitly modeling reactant-to-product transformations. Here we…
Pierfrancesco Ombrini, Qidi Wang, Alexandros Vasileiadis, Fangting Wu + 6 more
Effective optimization and control of lithium-ion batteries cannot neglect the relation between fundamental physicochemical phenomena and performance. In this work, we apply a multi-step charging protocol to commercially relevant electrodes, such as LiNi0.8Mn0.1Co0.1O2 (NMC811), LiFePO4 (LFP), LiMn1.5Ni0.5O4 (LMNO)…
Beau Sievers, Ida Momennejad
We present the Spreading Activation and Memory PLasticity Model (SAMPL), a computational model of how memory retrieval changes memories. SAMPL restructures memory networks as a function of spreading activation and plasticity. Memory networks are represented as graphs of items in which edge weights capture the strength…
Gavin Chiu, Sam J Gilbert
Intention offloading involves using external reminders such as diaries, to-do lists, and digital alerts to help us remember delayed intentions. Recent studies have provided evidence for various cognitive and metacognitive factors that guide intention offloading, but little research has investigated the physical cost of…
Authors not listed
Large Language Models (LLMs) based on transformer architectures excel at internet-scale tasks. However, real-world scientific scenarios—such as synthetic chemistry laboratories and autonomous experimental setups—typically involve incremental data generation in batches as new chemical reactions are conducted, unlike…
Authors not listed
This paper develops a comprehensive theoretical framework for designing quantum memory systems with enhanced resilience to thermal decoherence through engineered lattice geometries and protective structures. We formulate a unified mathematical description connecting material properties, geometric configurations, and…
Emilio Salinas, Bashirul I Sheikh
Our ability to recall details from a remembered image depends on a single mechanism that is engaged from the very moment the image disappears from view.
Tobore Onojighofia Tobore
Understanding the mechanisms behind memory, learning, and behavior is crucial to human development and significant research has been done in this area. Classical and operant conditioning and other theories of learning have elucidated different mechanisms of learning and how it modulates behavior. Even with advances in…
Bastian Wiederhold
As anyone who has tried to memorize a one-hundred-digit number can attest, acquisition of numerical information typically proceeds at less than one bit/s. If human memory operated at this speed in general, even a simple conversation would not be possible. Indeed, through techniques such as the memory palace, which…
Roselyne J. Chauvin, Annie Zheng, Athanasia Metoki, Samuel R. Krimmel + 19 more
Memory athletes can achieve superior performance (e.g., memorizing 339 digits in 5 minutes) with extensive daily training, by converting abstract information into vivid scenes, and placing them along a mental path, that is then retraced at recall (Method of Loci). Understanding the brain mechanisms underlying such…