Adaptive behavior with stable synapses
Abstract: Behavioral changes in animals and humans, as a consequence of an error or a verbal instruction, can be extremely rapid. Improvement in behavioral performances are usually associated in machine learning and reinforcement learning to synaptic plasticity, and, in general, to changes and optimization of network parameters. However, such rapid changes are not coherent with the timescales of synaptic plasticity, suggesting that the mechanism responsible for that could be a dynamical network reconfiguration. In the last few years, similar capabilities have been observed in transformers, foundational architecture in the field of machine learning that are widely used in applications such as natural language and image processing. Transformers are capable of in-context learning, the ability to adapt and acquire new information dynamically within the context of the task or environment they are currently engaged in, without the need for significant changes to their underlying parameters. Building upon the notion of something unique within transformers enabling the emergence of this property, we claim that it could also be supported by input segregation and dendritic amplification, features extensively observed in biological networks. We propose an architecture composed of gain-modulated recurrent networks that excels at in-context learning, showing abilities inaccessible to standard networks.
- Contextual behavior and neural circuits. Frontiers in neural circuits, 7:84, 2013.
- Matching networks for one shot learning. Advances in neural information processing systems, 29, 2016.
- An embarrassingly simple approach to zero-shot learning. In International conference on machine learning, pages 2152–2161. PMLR, 2015.
- Language models are few-shot learners. Advances in neural information processing systems, 33:1877–1901, 2020.
- The transient nature of emergent in-context learning in transformers. Advances in Neural Information Processing Systems, 36, 2024.
- Attention is all you need. Advances in neural information processing systems, 30, 2017.
- Gpt-4 technical report. arXiv preprint arXiv:2303.08774, 2023.
- Zero-shot text-to-image generation. In International conference on machine learning, pages 8821–8831. Pmlr, 2021.
- Hopfield networks is all you need. arXiv preprint arXiv:2008.02217, 2020.
- In-context learning and induction heads. arXiv preprint arXiv:2209.11895, 2022.
- Transformers learn in-context by gradient descent. In International Conference on Machine Learning, pages 35151–35174. PMLR, 2023.
- Efficiently modeling long sequences with structured state spaces. In International Conference on Learning Representations, 2021.
- Resurrecting recurrent neural networks for long sequences. In International Conference on Machine Learning, pages 26670–26698. PMLR, 2023.
- Gated recurrent neural networks discover attention. arXiv preprint arXiv:2309.01775, 2023.
- Illuminating dendritic function with computational models. Nature Reviews Neuroscience, 21(6):303–321, 2020.
- Matthew Larkum. A cellular mechanism for cortical associations: an organizing principle for the cerebral cortex. Trends in neurosciences, 36(3):141–151, 2013.
- Learning by the dendritic prediction of somatic spiking. Neuron, 81(3):521–528, 2014.
- Towards deep learning with segregated dendrites. eLife, 6:e22901, 2017.
- Dendritic cortical microcircuits approximate the backpropagation algorithm. In S. Bengio, H. Wallach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, and R. Garnett, editors, Advances in Neural Information Processing Systems 31, pages 8721–8732. Curran Associates, Inc., 2018.
- Burst-dependent synaptic plasticity can coordinate learning in hierarchical circuits. Nature neuroscience, 24(7):1010–1019, 2021.
- Beyond spiking networks: The computational advantages of dendritic amplification and input segregation. Proceedings of the National Academy of Sciences, 120(49):e2220743120, 2023.
- Matthew Larkum. A cellular mechanism for cortical associations: an organizing principle for the cerebral cortex. Trends in Neurosciences, 36(3):141 – 151, 2013.
- A solution to the learning dilemma for recurrent networks of spiking neurons. Nature communications, 11(1):1–15, 2020.
- Towards biologically plausible dreaming and planning. arXiv preprint arXiv:2205.10044, 2022.
- In-context reinforcement learning with algorithm distillation. arXiv preprint arXiv:2210.14215, 2022.
- Excitatory and inhibitory interactions in localized populations of model neurons. Biophysical journal, 12(1):1–24, 1972.
- Re-visiting the echo state property. Neural networks, 35:1–9, 2012.
- Extreme learning machine: theory and applications. Neurocomputing, 70(1-3):489–501, 2006.
- Physiology of layer 5 pyramidal neurons in mouse primary visual cortex: coincidence detection through bursting. PLoS computational biology, 11(3):e1004090, 2015.
- Robert K Heaton. Wisconsin card sorting test. Psychological assessment resources, 1981.
- Abstract representations emerge in human hippocampal neurons during inference behavior. bioRxiv, pages 2023–11, 2023.
- Bandit problems: sequential allocation of experiments (monographs on statistics and applied probability). London: Chapman and Hall, 5(71-87):7–7, 1985.
- Richard GM Morris. Spatial localization does not require the presence of local cues. Learning and motivation, 12(2):239–260, 1981.
- Pre-train, prompt, and predict: A systematic survey of prompting methods in natural language processing. ACM Computing Surveys, 55(9):1–35, 2023.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.