BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//ARNI - ECPv6.18.4//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-ORIGINAL-URL:https://arni-institute.org
X-WR-CALDESC:Events for ARNI
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20250309T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20251102T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20260308T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20261101T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20270314T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20271107T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20261007T130000
DTEND;TZID=America/New_York:20261007T140000
DTSTAMP:20260930T174603Z
CREATED:20260930T174603Z
LAST-MODIFIED:20260930T174603Z
UID:2667-1791378000-1791381600@arni-institute.org
SUMMARY:ARNI Continual Learning Working Group Meeting
DESCRIPTION:Continuation of prior meetings. \nZoom: upon request @ arni@columbia.edu
URL:https://arni-institute.org/event/arni-continual-learning-working-group-meeting-6/
LOCATION:Schapiro 416
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20261014T130000
DTEND;TZID=America/New_York:20261014T140000
DTSTAMP:20261013T142712Z
CREATED:20261013T142712Z
LAST-MODIFIED:20261013T142712Z
UID:2738-1791982800-1791986400@arni-institute.org
SUMMARY:ARNI Continual Learning Working Group Meeting
DESCRIPTION:Continuation of prior meetings \nZoom: Upon request @arni@columbia.edu
URL:https://arni-institute.org/event/arni-continual-learning-working-group-meeting-7/
LOCATION:CBS 453
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20261019T150000
DTEND;TZID=America/New_York:20261019T160000
DTSTAMP:20261008T163210Z
CREATED:20260930T172828Z
LAST-MODIFIED:20261008T163210Z
UID:2660-1792422000-1792425600@arni-institute.org
SUMMARY:Speaker:Divyansha Lachi (UPenn) and Thorir Ingolfsson (ETH)– ARNI Frontier Models for Neuroscience and Behavior Working Group
DESCRIPTION:Divyansha’s talk\nTitle: Scaling Neural Motor Decoding via Decoupled Behavioral Pretraining\nAbstract:\nRecent neural decoding approaches achieve strong performance by training on large collections of paired neural-behavioral recordings\, but such datasets are expensive\, invasive\, and difficult to scale. In contrast\, behavioral data is abundant and easy to collect\, raising the question of whether neural decoding can benefit from decoupling behavioral and neural representation learning. We introduce BeeMO\, a flexible multi-session framework that enables training with arbitrary mixtures of paired neural-behavioral recordings and unpaired behavioral data. A behavior decoder learns movement dynamics directly from behavioral trajectories\, while neural activity is incorporated through cross-attention layers when available\, allowing behavioral and neural supervision to scale independently. By decoupling neural and behavioral data\, we unlock a new dimension for scaling: unpaired behavioral data\, which is substantially cheaper and easier to collect than paired neural recordings\, can be used directly to improve decoding performance. Across multiple intracortical motor datasets and tasks in nonhuman primates\, we show that incorporating unpaired behavioral data consistently improves decoding performance\, particularly […]\, and that incorporating task-aligned behavioral trajectories can further improve transfer. We further show that behavior-only pretraining\, without any neural pretraining\, outperforms single-session supervised baselines and is comparable to models pretrained on substantially larger neural datasets. Together\, our results suggest that scaling behavioral data offers a practical and cost-effective path toward neural decoding models that generalize across subjects and tasks. \nThorir’s talk\nTitle: Foundation Models for Wearables: From EEG and EMG to Multimodal Models at the Edge\nAbstract:\nWearable devices can continuously record EEG\, EMG\, ECG\, and PPG\, but labels are scarce\, electrode layouts vary across devices\, and inference must fit within a power budget of tens of milliwatts. This talk gives an overview of our work at ETH Zürich on foundation models built under these constraints. I will start with EEG: LUNA\, a topology-agnostic model that handles arbitrary electrode montages\, and FEMBA\, a state-space alternative to transformers that scales linearly with sequence length. I will then turn to the pre-training objective\, comparing masked reconstruction with latent-space prediction (LeJEPA) in LuMamba and examining what each learns. Next\, I will present TinyMyo\, a compact EMG foundation model that transfers across hand gesture classification\, hand kinematic regression\, and speech production and recognition\, and runs on an ultra-low-power microcontroller. Finally\, PanLUNA extends the approach to joint modeling of EEG\, ECG\, and PPG. \nZoom: Upon request @ arni@columbia.edu
URL:https://arni-institute.org/event/speakerdivyansha-lachi-upenn-and-thorir-ingolfsson-eth-arni-frontier-models-for-neuroscience-and-behavior-working-group/
LOCATION:Virtual
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20261020T020000
DTEND;TZID=America/New_York:20261020T160000
DTSTAMP:20261013T131325Z
CREATED:20260930T173946Z
LAST-MODIFIED:20261013T131325Z
UID:2665-1792461600-1792512000@arni-institute.org
SUMMARY:Speaker: Joshua Gold ARNI WG Multi-resource-cost optimization of neural network models
DESCRIPTION:Joshua Gold \nTitle: Flexible information bottlenecks in the brain \nAbstract: \nThe brain must make sense of the world from a constant bombardment of incoming information but using limited computational and other resources. These limitations introduce information bottlenecks in the brain\, where the information that is processed is a compressed (capacity-limited) version of the information that is received. I will describe ongoing projects in my lab that aim to better understand the nature of these bottlenecks\, in particular their flexibility across conditions and individuals. The first part of the talk will focus on recent experimental findings showing that sensory adaptation\, a form of bottleneck evident in the activity of single neurons throughout the brain\, can adjust flexibly to recent stimulus statistics to support effective decision-making in dynamic environments. The second part of the talk will focus on recent theoretical work\, supported by human psychophysics\, that aims to account for this kind of flexibility in terms of balancing inherent trade-offs between the information capacity and accuracy of processes that rely on making inferences about the world from noisy observations. This work also provides new perspectives on optimal behavior\, which can be thought of in terms the efficient and effective use of capacity-limited information. \nZoom: Upon request @ arni@columbia.edu
URL:https://arni-institute.org/event/speaker-joshua-gold-arni-wg-multi-resource-cost-optimization-of-neural-network-models/
LOCATION:Virtual
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20261111T130000
DTEND;TZID=America/New_York:20261111T140000
DTSTAMP:20261111T181047Z
CREATED:20261111T181047Z
LAST-MODIFIED:20261111T181047Z
UID:2765-1794402000-1794405600@arni-institute.org
SUMMARY:ARNI Continual Learning Working Group
DESCRIPTION:The next meeting of our working group will be tomorrow\, Wed Nov 11 at 1pm on Zoom and in person at Schermerhorn 406 at Columbia. We will feature a presentation by Hussein Zolfaghari who will talk about Continual Learning for Personalized Assistive Control: An OpenSim Exosuit Case Study. The abstract is detailed below.\n\nContinual Learning for Personalized Assistive Control: An OpenSim Exosuit Case Study\n \nMany individuals experience reduced muscle strength due to injury\, aging\, or other physical conditions\, motivating assistive technologies that can adapt to users with different and changing physical capabilities. Such changes also provide a natural continual learning problem: an intelligent system must adapt to new user conditions while retaining knowledge acquired under previous conditions. This study investigates a neural network controller with replay-based continual learning for personalized elbow assistance under changing muscle-strength conditions. A cable-driven elbow exosuit was simulated using an OpenSim upper-limb model with a 5-kg payload. The neural controller was trained sequentially to generate assistive cable forces for elbow trajectory tracking under four simulated muscle-strength levels corresponding to 90%\, 60%\, 45%\, and 30% of healthy reference capacity. The controller was subsequently evaluated on previously learned conditions to quantify catastrophic forgetting. Standard sequential training showed substantial degradation on earlier conditions\, whereas replay-based continual learning largely preserved previously learned control performance. Across three selected validation comparisons\, continual learning reduced the mean elbow-angle tracking RMSE from 3.81° to 0.048°\, corresponding to a 98.8% reduction. Beyond this specific control application\, physics-based musculoskeletal simulation offers a rich environment for studying continual learning under naturally evolving human conditions. OpenSim can provide controlled changes in muscle capacity\, fatigue\, movement\, payload\, impairment\, and user characteristics\, creating opportunities to systematically study adaptation\, forgetting\, transfer\, and benchmarking of continual-learning algorithms in human-centered intelligent systems.
URL:https://arni-institute.org/event/arni-continual-learning-working-group-7/
LOCATION:Schermerhorn 406
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20261113T113000
DTEND;TZID=America/New_York:20261113T133000
DTSTAMP:20261111T173947Z
CREATED:20261111T173910Z
LAST-MODIFIED:20261111T173947Z
UID:2759-1794569400-1794576600@arni-institute.org
SUMMARY:CTN Speaker: Bing Wen Brunton
DESCRIPTION:Bing Wen Brunton\n\n\n\n\n\nSeminar Time: 11:30am\n\nDate: Fri 11/13/26\nSeminar Location: JLG\, L5-084 \n\n\nHost: Chase King\n\n \n\nTitle: Embodied intelligence: From connectomes to body models \nAbstract: Brains and nervous systems evolved jointly with the bodies they inhabit. To understand brains\, behavior\, and intelligence\, we must embrace the embodied nature of neural computation. In this talk\, I will describe some ongoing collaborative work towards putting dynamic models of brains as controllers inside the bodies they evolved to sense and actuate. I will present our recent work developing connectome simulations of the Drosophila ventral nerve cord (VNC)\, where we identified a three-neuron circuit as a putative central pattern generator (CPG) for fly walking. I suggest that integrating simulations of connectomes with physically realistic body models will allow us to investigate active behavioral sequences and how specific recurrent neural circuits support robust\, embodied intelligence in animals.
URL:https://arni-institute.org/event/2759/
LOCATION:Zuckerman Institute – L5-084\, 3227 Broadway\, New York\, NY\, United States
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20261120T083000
DTEND;TZID=America/New_York:20261120T170000
DTSTAMP:20261027T172648Z
CREATED:20261027T172611Z
LAST-MODIFIED:20261027T172648Z
UID:2742-1795163400-1795194000@arni-institute.org
SUMMARY:Continual Learning in Minds and Machines Workshop
DESCRIPTION:Overview \nWebpage\n\n\n\nHow can AI systems keep learning from experience? \n\n\n\n\nAn assistant working with a person or organization over time must remember past interactions\, learn from limited feedback\, and adapt as circumstances change. Continual learning studies how systems acquire new knowledge while preserving what remains useful. These questions connect AI research with the study of human learning and memory. \nCo-organized by Columbia Business School’s AI in Business Initiative and the NSF AI Institute for Artificial and Natural Intelligence (ARNI)\, this workshop brings together researchers in machine learning\, neuroscience\, and cognitive science with industry teams building continually learning systems. We will examine how minds and machines learn over time\, what biological memory can teach us about AI\, and how to measure whether accumulated experience improves an agent’s decisions. \nPlease note that this event is intended for researchers in the field\, not for Master’s or MBA students. A promotional code is required for registration. Please contact ARNI@columbia.edu if you are interested in registering for this workshop. 
URL:https://arni-institute.org/event/continual-learning-in-minds-and-machines-workshop/
LOCATION:Columbia Business School\, David Geffen Hall
ORGANIZER;CN="ARNI":MAILTO:arni@columbia.edu
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20261214T150000
DTEND;TZID=America/New_York:20261214T160000
DTSTAMP:20261118T184149Z
CREATED:20261118T184149Z
LAST-MODIFIED:20261118T184149Z
UID:2769-1797260400-1797264000@arni-institute.org
SUMMARY:Speaker: Eric Denovellis (UCSF)– ARNI Frontier Models for Neuroscience and Behavior Working Group
DESCRIPTION:Title and Abstract:TBD \nZoom Link: Upon request @ ARNI@columbia.edu
URL:https://arni-institute.org/event/speaker-eric-denovellis-ucsf-arni-frontier-models-for-neuroscience-and-behavior-working-group/
LOCATION:Virtual
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20261215T110000
DTEND;TZID=America/New_York:20261215T123000
DTSTAMP:20261118T185234Z
CREATED:20261118T185234Z
LAST-MODIFIED:20261118T185234Z
UID:2773-1797332400-1797337800@arni-institute.org
SUMMARY:ARNI Emerging Researchers (AER): Talk Series
DESCRIPTION:ARNI Emerging Researchers (AER) : Talk Series \nDate: 12/15 \nLocation: Zoom upon request @ arni@columbia.edu \nTime: 11am to 12:30pm  \nDetails:  \nTo build on the momentum from the Annual Retreat\, ARNI is relaunching its series of informal online mini-talks in AY27. Each session will feature 2–3 students or postdoctoral researchers presenting brief updates on their current projects\, followed by an open discussion. \nThese sessions will provide a welcoming and collaborative environment for participants to share their research\, receive feedback from peers across ARNI\, and exchange ideas. This space will also offer valuable opportunities for students and postdocs to connect\, network\, and explore potential research collaborations across the institute. \nSpeakers:  \nAmrut Nadgir\, Graduate student at UPenn \nDongrui Deng\, Graduate student at Carnegie Mellon  \nZeyuan Wang\, Postdoc at UPenn  \nTitle and Abstract: TBD  \n 
URL:https://arni-institute.org/event/arni-emerging-researchers-aer-talk-series/
LOCATION:Virtual
END:VEVENT
END:VCALENDAR