BEGIN:VCALENDAR METHOD:PUBLISH PRODID:Microsoft Exchange Server 2010 VERSION:2.0 X-WR-CALNAME:Calendar BEGIN:VTIMEZONE TZID:UTC BEGIN:STANDARD DTSTART:16010101T000000 TZOFFSETFROM:+0000 TZOFFSETTO:+0000 END:STANDARD BEGIN:DAYLIGHT DTSTART:16010101T000000 TZOFFSETFROM:+0000 TZOFFSETTO:+0000 END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT DESCRIPTION:Abstract: The boundary-boundary art-gallery problem asks\, give n a polygon $P$ representing an art-gallery\, for a minimal set of guards that can see the entire boundary of $P$ (the wall of the art gallery)\, wh ere the guards must be placed on the boundary. We show that this art-galle ry variant is in NP. In order to prove this\, we develop a constraint-prop agation procedure for continuous constraint satisfaction problems where ea ch constraint involves at most 2 variables.\n\nThe X-Y variant of the art- gallery problem is the one where the guards must lie in X and need to see all of Y. Each of X and Y can be either the vertices of the polygon\, the boundary of the polygon\, or the entire polygon\, giving 9 different varia nts. Previously\, it was known that X-vertex and vertex-Y variants are all NP-complete and that the point-point\, point-boundary\, and boundary-poin t variants are $\\exists \\mathbb{R}$-complete [Abrahamsen\, Adamaszek\, a nd Miltzow\, JACM 2021][Stade\, SoCG 2025]. However\, the boundary-boundar y variant was only known to lie somewhere between NP and $\\exists \\mathb b{R}$.\n\nThe X-vertex and vertex-Y variants can be straightforwardly redu ced to discrete set-cover instances. In contrast\, we give example to show that a solution to an instance of the boundary-boundary art-gallery probl em sometimes requires placing guards at irrational coordinates\, so it unl ikely that the problem can be easily discretized.\n UID:040000008200E00074C5B7101A82E008000000002270BCC29FA1DC01000000000000000 010000000E36DB278FDB2254AA9F798714CAB77DA SUMMARY:NP-Membership for the Boundary-Boundary Art-Gallery Problem DTSTART;TZID=UTC:20260223T170000 DTEND;TZID=UTC:20260223T180000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:LSRC D344 X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:About the Event\n\nThe AiiCE Lecture Series presents Student Vo ices in Action: AiiCE Student Advisory Board\, a special event which highl ights student leadership in computing education. This special session will feature short talks from members of the 2025 - 2026 AiiCE Student Advisor y Board\, both undergraduate and graduate students. Attendees will have th e unique opportunity to hear directly from our students as they reflect on their projects\, the challenges and opportunities they’ve encountered\, and what it takes to build supportive communities in computing education. Three of the six student speakers are Duke Computer Science students.\n\ nModerator\n\nDr. Valerie Barr\, Margaret Hamilton Distinguished Professor of Computer Science at Bard College\n\nStudent Speakers\n\n2025 - 2026 Ai iCE Student Advisory Board\n\n * Bridget Agyare (University of Illinois Urbana-Champaign)\n * Ashita Birla (Duke University)\n * Ashlyn Cam pbell (University of Michigan)\n * Alveena Nadeem (Duke University)\n * Neha Shukla (Duke University)\n * Jessica Yauney (Stanford Universi ty)\n\nRegister\n\nhttps://duke.is/StudentVoices (Zoom registration is req uired)\n UID:040000008200E00074C5B7101A82E00800000000A25E7903AFA5DC01000000000000000 010000000C117C60C8AA85A45BB6C0C19E9D63666 SUMMARY:Student Voices in Action: AiiCE Student Advisory Board DTSTART;TZID=UTC:20260224T220000 DTEND;TZID=UTC:20260224T230000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:Virtual Event - Registration is required. X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:Lunch will be served at 11:45AM.\n\n\n\nBiography:\n\nDan Roth is the Chief AI Scientist at Oracle and the Eduardo D. Glandt Distinguishe d Professor at the University of Pennsylvania. Previously\, Dan was a VP/D istinguished Scientist at AWS AI where he led the scientific effort behind Amazon’s first-generation GenAI products\, including Titan Models\, Ama zon Q\, and Amazon Bedrock. Dan is a Fellow of the AAAS\, ACM\, AAAI\, an d ACL\, and a recipient of the IJCAI John McCarthy Award “for major conc eptual and theoretical advances in the modeling of natural language unders tanding\, machine learning\, and reasoning.” He has published broadly in natural language processing\, machine learning\, knowledge representation and reasoning\, and learning theory\, was the Editor-in-Chief of the Jour nal of Artificial Intelligence Research (JAIR) and has served as a Program Chair and Conference Chair for the major conferences in his research area s. Roth has been involved in several ML/NLP/GenAI startups in domains that range from legal and compliance to health care. Dan received his B.A Summ a cum laude in Mathematics from the Technion\, Israel and his Ph.D. in Com puter Science from Harvard University in 1995.\n\nAbstract:\n\nThe rapid p rogress made over the last few years in generating linguistically coherent natural language has blurred\, in the mind of many\, the difference betwe en natural language generation\, understanding\, knowledge retrieval and u se\, and the ability to reason with respect to the world. Nevertheless\, r eliably and consistently supporting high-level decisions that depend on na tural language understanding and heterogenous information retrieval is sti ll difficult for fundamental reasons that range from computational complex ity to data organization in the wild (and are here to stay).\n\nI will dis cuss some of the challenges underlying reasoning and information access\, argue that we should exploit what LLMs do well while delegating responsibi lity to special purpose models and solvers for decision making\, and prese nt some of our work in this space. I hope to collectively acknowledge some of the key GenAI myths and their consequences\, think about their underly ing causes\, and discuss ways to move forward.\n\n\n\nZoom Link: https://d uke.zoom.us/j/99101437873?pwd=FxnXqbrds6niuKXkQOaPA7JWhKUPJS.1\n\nMeeting ID: 991 0143 7873\n\nPasscode: 045260\n UID:040000008200E00074C5B7101A82E00800000000C3350CEC5999DC01000000000000000 010000000FC4BCC17B2AC314F94C4B3135AF598FC SUMMARY:On Reasoning & Retrieving LLMs: Myths\, Merits\, and How to Move Fo rward DTSTART;TZID=UTC:20260226T164500 DTEND;TZID=UTC:20260226T180000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:LSRC D106 X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:Abstract: In many decision-making scenarios\, individuals strat egically choose what information to disclose to optimize their own outcome s. It is unclear whether such strategic information disclosure can lead to good societal outcomes. To address this question\, we consider a competit ive Bayesian persuasion model in which multiple agents selectively disclos e information about their qualities to a principal\, who aims to choose th e candidates with the highest qualities. Using the price-of-anarchy framew ork\, we quantify the inefficiency of such strategic disclosure. We show t hat the price of anarchy is at most a constant when the agents have indepe ndent quality distributions\, even if their utility functions are heteroge neous. This result provides the first theoretical guarantee on the limits of inefficiency in Bayesian persuasion with competitive information disclo sure.\n UID:040000008200E00074C5B7101A82E0080000000060B922FE2AA7DC01000000000000000 010000000BC909A2ED2015A4C831AEEEBEB67E7E5 SUMMARY:The Price of Competitive Information Disclosure DTSTART;TZID=UTC:20260302T170000 DTEND;TZID=UTC:20260302T180000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:LSRC D344 X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:Lunch will be served at 12:45PM.\n\nBio\n\nDr. Fan Zhang is an Assistant Professor in the Department of Computer Science at Yale Universi ty. His research spans computer security and applied cryptography\, with a focus on novel systems with decentralized trust and governance. Several o f his works\, notably zkTLS\, confidential smart contracts\, trustless bri dges\, and anonymous communication\, have been adopted by the industry. Hi s research has been supported by awards and grants from the NSF\, the Ethe reum Foundation\, Flashbots\, Mysten Labs\, the Yale Roberts Innovation Aw ard\, and IBM. He received his Ph.D. in Computer Science from Cornell Univ ersity\, advised by Prof. Ari Juels\, and his Bachelor of Science from Tsi nghua University\, China.\n\nAbstract\n\nCryptography has long pursued the goal of building systems that can enforce commitments without relying on trusted intermediaries. Blockchain-based smart contracts represent one rea lization of this vision: their transparency and integrity make it difficul t for participants to violate agreed-upon rules. Yet much of today’s dig ital infrastructure—from advertising networks and social media platforms to online marketplaces—remains centralized and opaque. These systems op erate as black boxes\, leaving users and researchers with limited ability to verify how they behave in practice.\n\nIn this talk\, I explore how cry ptographic techniques can enable decentralized auditing of trusted platfor ms. I first discuss how tools such as zkTLS and trusted execution environm ents (TEEs) allow users to produce privacy-preserving and authentic eviden ce about what web services deliver to them\, strengthening community-drive n auditing efforts such as ad transparency studies. I then present Verifia ble Aggregate Receipts (VAR)\, a protocol that enables users to collective ly prove aggregate properties of platform behavior—such as the number of times content was served—without trusting the platform. Together\, thes e techniques illustrate a broader vision of reversing the digital panoptic on\, where users can cryptographically audit powerful online systems.\n\n\ n\nZoom Link: https://duke.zoom.us/j/97371787328?jst=3\n\n\n UID:040000008200E00074C5B7101A82E00800000000947D63502BB2DC01000000000000000 010000000650E7EEE28BDE74FA9A67204C5C9D88A SUMMARY:Reversing the Panopticon: Decentralized Auditing of Trusted Platfor ms DTSTART;TZID=UTC:20260318T170000 DTEND;TZID=UTC:20260318T180000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:LSRC D344 X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:Abstract\n\nConsider the task of sampling n independent bits\, each equal to 1 with probability 1/4. Given 2n uniform random bits as inpu t\, this distribution can be sampled exactly by letting each output bit be the AND of two input bits. What if we are given only 1.99n uniform random bits? Can we still sample the distribution so that each output depends on only constant many input bits?\n\nIn this talk\, I will present several r esults on sampling product distributions in a local and randomness-efficie nt manner:\n\n- Let D = (D_1\, D_2\, ...\, D_n) be a product distribution where each D_i has constant support and dyadic probabilities (i.e.\, proba bilities of the form (a/2^b)\, where a\,b are integers). Then D can be sam pled in constant time in the bit-probe model (equivalently\, in NC^0) usin g (h(D) + ε)n random bits\, where h() is the binary entropy function\, up to exponentially small statistical error. Moreover\, the dyadic assumpti on is necessary.\n\n- We characterize the tradeoffs between locality and s tatistical distance for sampling the 1/4-biased distribution with 1.99n un iform bits. With 2 bit probes\, essentially no nontrivial approximation is possible. With 3 bit probes\, we construct a sampler achieving statistica l distance 1/poly(n)\, and show this is best possible. Finally\, 4 bit pr obes suffice to achieve exponentially small statistical distance.\n\n- Eve ry p-biased distribution can be sampled in constant time in the cell-probe model with randomness complexity h(p)n + √n · polylog(n)\, up to a pol ynomially small statistical distance.\n\nOur constructions rely on pseudor andom distributions that are bounded-uniform on average. These are obtain ed using tools from low-density parity-check codes\, as well as recent res ults on succinct and retrieval data structures.\n UID:040000008200E00074C5B7101A82E00800000000801BCFC88DB8DC01000000000000000 010000000323BD5C04B586F41A9ED868C1C4C47DD SUMMARY:Local Samplers for Product Distributions DTSTART;TZID=UTC:20260323T160000 DTEND;TZID=UTC:20260323T170000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:LSRC D344 X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:Bio\n\nSaket Tiwari is a postdoctoral researcher at the Univers ity of California\, Santa Barbara\, in the REAL AI Lab. He received his Ph D from Brown University under the supervision of Prof. George Konidaris. H is work focuses on developing a new effective theory of deep RL in continu ous environments\, combining elements of optimal control\, geometry\, and optimization theory. Prior to his PhD\, he obtained a master’s degree fr om the University of Massachusetts Amherst and a bachelor’s degree from IIT Bombay.\n\nAbstract\n\nDeep reinforcement learning (RL) for control fr om pixels still relies heavily on discretized models\, where agents choose from finitely many actions and receive finitely many observations. This l imits our ability to design better optimizers and architectures\, and to u nderstand modern RL algorithms at a deeper level. In this talk\, I will in troduce and analyze a new optimizer for RL agents with high-dimensional ob servations and overparameterized neural network function approximators. To do this\, I study an episodic actor-critic algorithm for the continuous-t ime linear-quadratic regulator problem. The agent receives high-dimensiona l observations generated from low-dimensional latent states. To model over parameterized neural networks while keeping the analysis tractable\, we us e deep linear networks whose weight matrices have mutually orthonormal col umns\, meaning they lie on the Stiefel manifold. Under these assumptions\, we show that the learning dynamics of the high-dimensional problem is ana logous to optimization in the underlying low-dimensional latent state\, re vealing a hidden structure. We support this theory with experiments on pix el-based robotic control tasks trained end-to-end from reward\, without au xiliary representation losses\, showing that proximal policy optimization (PPO) with Stiefel-manifold optimization outperforms PPO with Adam. I will close by proposing a new metaphor for RL: one that is continuous in state s\, actions\, and time.\n\nZoom Link: https://duke.zoom.us/j/97206204162?p wd=zyUwHF2DZa4SMTm9N1zNvyE9rbuEcC.1\n UID:040000008200E00074C5B7101A82E00800000000079E70B167C1DC01000000000000000 0100000007AD549195768D5408767FF262F8AE105 SUMMARY:Structure in High-Dimensional Reinforcement Learning DTSTART;TZID=UTC:20260403T190000 DTEND;TZID=UTC:20260403T200000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:LSRC D344 X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:Bio\n\nDr. Huajie Shao is a tenure-track Assistant Professor of Computer Science at William & Mary. Before that\, he obtained his Ph.D. d egree in Computer Science from the University of Illinois at Urbana-Champa ign (UIUC) in 2021. Dr. Shao’s research focuses on the intersection of m achine learning/foundation models\, control\, and embodied AI systems. To date\, he has published more than 60 papers in top-tier conferences and jo urnals\, including ICML\, ICLR\, KDD\, CVPR\, ACL\, TPAMI\, WWW\, and SIGI R. His work has received multiple best paper awards\, including the Best P aper Awards at KDD 2024\, ACM/IEEE CHASE 2025\, CAHSE 2024\, SenSys 2020\, and ICCPS 2017\, as well as the FUSION 2019 Student Paper Award and the U biComp 2019 Distinguished Paper Award.\n\n\n\nAbstract\n\nEmbodied AI\, su ch as autonomous vehicles and robotic systems\, has transformed everyday l ife\, reshaping how we live\, work\, and interact with the world. However\ , a key challenge remains: data-driven machine learning models often strug gle to generalize to unknown and dynamic environments. In this talk\, I wi ll introduce a series of physics-informed machine learning models that inc orporate physical laws into model design to enhance generalization in unse en environments. I will begin by presenting a generalizable\, physics-info rmed state-space model for embodied AI systems based on partially known ph ysics knowledge. I will then discuss WestWorld\, a knowledge-encoded traje ctory world model for multi-robot learning and control. The proposed model demonstrates superior performance in zero- and few-shot trajectory predic tion\, as well as in downstream model-based control across various robotic s. Finally\, I will conclude by outlining potential future directions for advancing world models in embodied AI.\n UID:040000008200E00074C5B7101A82E008000000001A4A580894C6DC01000000000000000 010000000A1D042878EF4B3448E5D98CF2169527C SUMMARY:Physics-Guided Machine Learning for Embodied AI DTSTART;TZID=UTC:20260409T160000 DTEND;TZID=UTC:20260409T170000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:LSRC D344 X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:Lunch will be served at 11:45AM.\n\nBio:\n\nDr. Stephen Ni-Hahn is a Postdoctoral Associate at Duke University\, and holds a PhD in Elect rical and Computer Engineering from Duke University. His research investig ates generative AI for computational creativity\, specifically on incorpor ating domain knowledge to enable greater interpretability and controllabil ity in AI music systems. His work has been featured and published broadly in many top machine learning and music conferences such as NeurIPS\, AAAI\ , KDD\, ISMIR\, and SMT.\n\nAbstract:\n\nThe intersection of artificial in telligence and music is vast and quickly evolving\, with models such as Su no enabling amateur music enthusiasts to generate impressive compositions with a simple text prompt. There is further an urgent need for AI systems that humans can control. Unfortunately\, the vast majority of recent syste ms rely on black box models trained on massive datasets\, heavily limiting themselves based on available data. Furthermore\, most models are expecte d to learn complex notions like music-theoretical structure without any gu idance\, leading to poor consistency and a lack of large-scale structure. My research addresses these limitations by integrating domain knowledge fr om music theory to enhance interpretability and human controllability for AI music generation\, enabling the generation of coherent and enjoyable mu sic that outperforms much larger state-of-the-art deep learning models. In this talk\, I will focus on three methods that facilitate this integratio n: SchenkComposer\, a theory-based framework for hierarchical melody gener ation\; AutoSchA\, which build on recent developments in Graph Neural Netw orks\, enabling AI models to interpret deeper musical connections in a mor e human way\; and E-Motion Baton\, a real-time human-in-the-loop conductin g simulator that incorporates human emotion and gesture for responsive mus ic generation.\n UID:040000008200E00074C5B7101A82E00800000000BB03E6B9F9AFDC01000000000000000 010000000A8392FE5A4C48343BAAC977EC29052A3 SUMMARY:Generative AI Meets Music Theory DTSTART;TZID=UTC:20260410T160000 DTEND;TZID=UTC:20260410T170000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:LSRC D106 X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:Please RSVP for lunch.\n\nCALLING RISING SOPHOMORES AND RISING CS JUNIORS and SENIORS!\n\nJoi n us for an information session on the Duke CS 4+1 program\, an accelerate d pathway for CS undergrads to earn both your bachelor’s and master’s degrees in computer science in only 5 years. Learn how the program works\, what makes a strong application\, and how it can help you advance your ac ademic and career goals.\n\nLunch will be provided--come eat\, ask questio ns\, and explore your next step! We look forward to seeing you there.\n\nP lease RSVP here so that we cater enough food: https://forms.cloud.microsof t/r/UVZkYnF8va [authentication required]\n\nFor more info contact Josh Boy d at j.boyd@duke.edu\n UID:040000008200E00074C5B7101A82E0080000000071E6004650C8DC01000000000000000 0100000006A47A60CE181034B87933F0343BFCAA5 SUMMARY:CS 4+1 Information Session DTSTART;TZID=UTC:20260415T160000 DTEND;TZID=UTC:20260415T170000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:LSRC D106 X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:Lunch will be served at 11:45AM.\n\nBio\nVyas Sekar is the Tan Family Professor of Electrical and Computer Engineering\nat Carnegie Mello n University. He is a member of CMU CyLab\, where he co-directs\nthe Futur e of Enterprise Security Initiative. He is also the co-founder and\nChief Technologist at Rockfish Data\, and the Chief Scientist at Conviva. His\nr esearch sits at the intersection of networks\, systems\, and security. His work\nhas been recognized with the SIGCOMM Rising Star Award\, the NSA Sc ience of\nSecurity prize\, the Intel Outstanding Researcher Award\, the SI GCOMM Test of\nTime Award\, and multiple Best Paper awards.\n\nAbstract\nA s Dickens famously wrote\, "It was the best of times\, it was the worst of \ntimes." We are entering an era where the ability to develop and deploy\n cloud-scale applications has been democratized\, yet our capacity to obser ve\,\nmanage\, and troubleshoot these environments has far outpaced human cognition.\n\nIn this landscape\, AIOps—the automation of systems manage ment—is critical for\ntransitioning from "human timescales" to the "mach ine timescales" at which\nmodern infrastructure operates. However\, rather than blindly applying\n"Artificial Intelligence" to the problem\, we argu e for a fundamental shift:\nreframing the "AI" in AIOps to focus on Action able Insights.\n\nIn this talk\, we identify the foundational requirements for AIOps in\ninternet-scale applications—Expressivity\, Efficiency\, E xplainability\, and\nEffort (the "4Es")—and present the design principle s necessary to build the\nnext generation of these systems. Finally\, we w ill share illustrative examples\nfrom academic and industry collaborations that embody these principles to\nachieve the 4E goals.\n UID:040000008200E00074C5B7101A82E00800000000A85BA77D0BC1DC01000000000000000 010000000FA967FDDE4F8B646BF79108C365961BC SUMMARY:Putting the AI in AIOps: From Artificial Intelligence to Actionable Insights DTSTART;TZID=UTC:20260417T160000 DTEND;TZID=UTC:20260417T170000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:LSRC D106 X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:🍦 Stay Cool with the CSSU!\n\nDuke CS\, are you Loco for Loc opops? The CSSU is helping you beat the heat with a sweet treat! Join us f or a pop-up giveaway to celebrate the spring weather. Whether you're a fru it fan or a cream enthusiast\, we've got you covered.\n\n * When: Tuesd ay\, April 21st\n * Time: 11:30 AM – 2:30 PM\n * Where: Look for t he CSSU table!\n UID:040000008200E00074C5B7101A82E00800000000291AA73791CEDC01000000000000000 0100000003C35B058C935AE439944AA88C390797F SUMMARY:Computer Science Student Union- Locopop Event DTSTART;TZID=UTC:20260421T153000 DTEND;TZID=UTC:20260421T183000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:Bryan Center Plaza X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:Lunch will be provided at 11:45am.\n\nAbstract:\n\nModern compu ting systems—from web services to large-scale AI—depend on cloud infra structure to provide performance\, scalability\, and isolation. Despite it s importance\, much of this infrastructure is built on software that is ul timately taken on trust — even as it is increasingly written using AI. W e test it\, patch it\, and rely on it — but we rarely have strong guaran tees that it is actually correct.\n\nIn this talk\, I will show how we can make this foundation more secure on Arm-based platforms\, which are incre asingly deployed end-to-end from mobile devices to the cloud. By reducing the trusted computing base\, formally verifying key security properties\, and automating that process so it scales to real system software\, we can begin to replace trust with machine-checked guarantees.\n\nBio:\n\nJason N ieh is Professor of Computer Science and Co-Director of the Software Syste ms Laboratory at Columbia University. Technologies he developed are widel y used in major operating system platforms\, including Android and Linux\, the largest cloud infrastructure providers\, including Amazon Web Service s and Google Cloud\, and Arm processors\, billions of which ship each year . He has published over a hundred peer-reviewed papers\, several of which have received best paper and test of time awards\, from MobiCom\, OSDI\, SIGCSE\, SIGMETRICS\, and SOSP. Nieh is a Fellow of the AAAS\, ACM\, IEEE \, and John Simon Guggenheim Memorial Foundation. He earned his B.S. from MIT and his M.S. and Ph.D. from Stanford University\, all in Electrical En gineering.\n UID:040000008200E00074C5B7101A82E00800000000968F273C48CBDC01000000000000000 0100000007A50C37E336E654F889A72FC31F505D3 SUMMARY:Securing the Foundations of the Cloud DTSTART;TZID=UTC:20260423T160000 DTEND;TZID=UTC:20260423T170000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:LSRC D106 X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT BEGIN:VEVENT DESCRIPTION:The 2026 Computer Science Commencement Ceremony will be held on Duke University’s campus at the Cameron Indoor Stadium (115 Whitford Dr ive\, Durham 27705).\n\nRSVP\n\nONLY graduating students who plan on parti cipating in the Computer Science commencement ceremony are required to reg ister. It is criti cal that graduating students let our organizers know they will take part i n the ceremony. Graduating students: Register by Friday\, May 1\, 2026.\n\ nNote to our graduating CS Second Majors - You are also welcome to partici pate in the Computer Science Departmental graduation ceremony--i.e.\, atte nd\, walk across the stage\, and have your picture taken with the CS Depar tment Chair while holding a diploma jacket. However\, you must register by May 1\, 2026 in order to participate. You will still need to see your fir st major department to receive your diploma cover.\n\nAll other guest stud ents\, faculty\, staff\, family and friends are welcome to attend--no regi stration or tickets required.\n\nRegistration form\n\n________________________________\nDuke Department of Computer Science 2026 Commencement Ceremony LIVE\n\nOur gra duation ceremony will be livestreamed on the Duke Computer Science YouTube Channel Friday\, May 8th at 1pm. Guests may click on the following link t o attend virtually:\n\nhttps://youtube.com/live/777yeHqNfmg?feature=share\ n\n________________________________\nGraduation Rehearsal\n\nWe will have a rehearsal of the Computer Science ceremony at Cameron Indoor Stadium bef ore actual commencement day at 2pm on Thursday\, May 7th.\n\nNOTE: While n ot required\, we highly recommend all graduating students who intend to pa rticipate in our departmental commencement exercises on Friday come to thi s practice session--to familiarize themselves with the planned program and the space.\n\n________________________________\nNEW! - Undergraduate Dipl oma Distribution\n\nThis year Trinity College of Arts and Sciences and Pra tt School of Engineering have decided to change the way that undergraduate students receive their diplomas. Undergraduate diplomas will NOT be hande d out during this year’s departmental ceremony\, unlike in past years. I nstead\, students will be given the empty diploma folder at the event. Lat er\, after commencement weekend\, the Registrar will mail the physical doc ument to undergraduate students.\n\nWhile all CS majors\, both 1st and 2nd majors are welcome to participate in the ceremony. Only CS 1st majors wil l be given their empty diploma folders at the CS departmental graduation. CS 2nd majors will need to get their diploma folders from their first majo r department or program.\n\nPLEASE NOTE: Graduate students will receive th eir diplomas via USPS as has been done customarily in the past. Graduate s tudents will not be provided with diploma covers.\n\n_____________________ ___________\nNEW! - Security Screening\n\nAll persons – faculty/staff\, graduating students\, and guests (including children) – are subject to s creening before entering Cameron Indoor Stadium. Children who are able to walk independently will also be required to walk through the security gate . Infants and small children\, however\, may be carried through by an olde r person. Walk-through metal detectors will be located at each entrance. I ndividuals will not need to remove small items from their person or small bag when passing through the metal detectors--for example\, cell phones\, keys\, cameras\, jackets\, belts\, coins\, wallets\, jewelry\, shoes\, etc . People entering should carry all items through the metal detector. Event Staff will inspect all personal items\, bags\, and equipment prior to ind ividuals entering the defined venue perimeter.\n\nAccessibility for Wheelc hairs and Assistive Devices\n\nThe walk-through metal detector lanes are n ot accessible for wheelchairs and other assistive devices. There will be d esignated Medical/Family Lanes present at each entrance accessible for al l individuals with disabilities. People utilizing these lanes will instead be screened using a hand-held metal detector with physical pat-downs as a final alternative option.\n\nMedical Devices and Physician's Advice\n\nIM PORTANT NOTE: Individuals with medical devices or other medical conditions should consult their doctor or device manufacturer if unsure of their abi lity to utilize a walk-through metal detector. Event staff will NOT give a dvice regarding any medical devices.\n\nAny person who does not consent to screening will be denied entry into the stadium.\n\nFor more details and additional questions\, please see Duke University's complete operations an d security policies for Cameron Indoor Stadium below.\n\nDownload Cameron_ Entry_Operations_and_Policy (pdf - 184.26 KB)\ n________________________________\nGuest Seating\n\nDoors open to guests a t 12noon. We ask that guests be seated no later than 12:30pm. Please allow plenty of time for travel and parking.\n\nAccessible Seating\n\nDesignate d accessible seating will be available at the venue to those guests using wheelchairs and/or persons with mobility\, visual\, or hearing impairments . A maximum of 2 companions may accompany each guest with a disability in these designated areas. Graduating students must register in advance for these special accom modations when RSVP’ing their participation in the event.\n\n___________ _____________________\nParking\n\nGeneral parking will be available at the Science Drive Garage located at 302 Science Drive\, Durham\, NC 27708. Cl ick here to see parking location on Google Maps.\n\nEnter the parking deck off Science Drive near JB Duke Hotel.\n\nRate: General parking is FREE.\n\nClearance: 8' 2"\n\nNOTE: Ge neral parking on campus is on a first-come basis\, so please allow plenty of time for parking on commencement day and throughout the weekend.\n\nADA Parking\n\nProximity parking is available at Cameron Indoor Stadium for t hose guests with mobility issues or other impairments which are impacted b y walking. Graduating students must register in advance for these special accommodations whe n RSVP’ing their participation in the graduation event.\n\n_____________ ___________________\nBags and Other Venue Policies\n\nBags and items large r than 12” x 12” x 6” (including backpacks\, duffel bags\, large pur ses\, suitcases\, wrapped gifts\, banners\, flags\, signs\, etc.) are proh ibited from entering the venue. Please leave such items in personal vehicl es or hotel rooms. Exceptions: bags for medical or childcare reasons.\n\nU mbrellas are permitted.\n\nNo pets allowed.\n\nProhibited items:\n\n * NO BALLOONS\, NO CONFETTI\, NO GLITTER\n * No Alcoholic Beverages\n * No Drones and No Unmanned Aircraft Systems (UAS) – https://drones.duke .edu/policies\n * No Backpacks\n * No Balloons\n * No Confetti\n * No Duffel Bags\n * No Flags on poles\n * No Glitter\n * No H azardous Materials\n * No Items obstructing the sightlines of other gue sts (including signs\, banners\, flags\, posters\, etc.)\n * No Laser P ointers\n * No Noisemakers that can disrupt the ceremony (air horns\, b ullhorns\, etc.)\n * No Pets (exception: service animals)\n * No Sui tcases\n * No Weapons of any kind (including pocket knives)\n * ...A nd again NO BALLOONS\, NO CONFETTI\, NO GLITTER\n\nTobacco-free Campus\n\n Duke University is a Tobacco-Free Campus. The use of any tobacco-based pro duct is prohibited\, including but not limited to cigarettes\, cigars\, ci garillos\, chewing tobacco\, snuff\, and electronic smoking devices (inclu ding e-cigarettes\, vaping products\, and IQOS). For more information on D uke University’s Tobacco-Free Campus policy\, please visit: healthy.duke .edu/tobaccofree/\n\n______________ __________________\nCS Commencement Schedule\n\n12 NOON Graduates arrive. \n12:45 PM Procession starts.\n1:00 PM Computer Science Diploma Cerem ony will begin at 1pm sharp.\n\n________________________________\nRegistra tion\n\nGraduating students please be sure to register your participation in the 2026 Duke D epartment of Computer Science Graduation Ceremony by Friday\, May 1\, 2026 .\n\n________________________________\nQuestions\n\nPlease contact Pam Spe ncer.\n UID:040000008200E00074C5B7101A82E00800000000376639ECD8A5DC01000000000000000 010000000AAC994BF9EF8704BBAC032A4EEEF082F SUMMARY:Computer Science 2026 Commencement DTSTART;TZID=UTC:20260508T170000 DTEND;TZID=UTC:20260508T190000 CLASS:PUBLIC PRIORITY:5 DTSTAMP:20260517T065138Z TRANSP:OPAQUE STATUS:CONFIRMED SEQUENCE:0 LOCATION:Cameron Indoor Stadium X-MICROSOFT-CDO-APPT-SEQUENCE:0 X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-INTENDEDSTATUS:BUSY X-MICROSOFT-CDO-ALLDAYEVENT:FALSE X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-CDO-INSTTYPE:0 X-MICROSOFT-DONOTFORWARDMEETING:FALSE X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MICROSOFT-REQUESTEDATTENDANCEMODE:DEFAULT X-MICROSOFT-ISRESPONSEREQUESTED:FALSE END:VEVENT END:VCALENDAR