{"id":2353,"date":"2025-09-24T18:30:35","date_gmt":"2025-09-24T18:30:35","guid":{"rendered":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/?page_id=2353"},"modified":"2025-09-24T18:53:09","modified_gmt":"2025-09-24T18:53:09","slug":"lineup","status":"publish","type":"page","link":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/","title":{"rendered":"Speaker Lineup with Talk Titles and Abstracts"},"content":{"rendered":"\n
\n
\n <\/span>
\n
\n \n
\n
\n

Speaker Lineup with Talk Titles and Abstracts<\/h1> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n
\n
\n
\n \n
\n
\n \n \"Kristian <\/a>\n \n <\/div>\n \n \n
\n

\n \n Dr. Anima Anandkumar <\/a>\n <\/h3>\n
\n Bren Professor at CalTech <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n

Neural Operators for AI+Science: Pushing the Frontiers of Scientific Discovery<\/h3>\n

Keynote<\/h5>\n
The main bottleneck in doing scientific research is the need for physical experiments in many areas. This means risky ideas are often discarded and the hypothesis space is traditionally restricted to regions of prior success. AI is disrupting this status quo by enabling physically-valid digital twins that reduce or even completely remove the need for physical experiments. AI models are orders of magnitude faster than traditional simulations, and often more accurate, since they can directly adapt to experimental and observational data. Since AI models are differentiable, they can be directly used for inverse design, enabling exploration and design optimization subject to diverse constraints, that was not possible before. Neural Operators enable multiscale and physics-informed learning for achieving high fidelity and training data efficiency in many areas. They have been successfully applied in weather and climate modeling, plasma evolution in nuclear fusion, designing novel medical devices and enabling autonomous flights under turbulence.<\/div> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n
\n
\n
\n \n
\n
\n \n \"Kristian <\/a>\n \n <\/div>\n \n \n
\n

\n \n Dr. Anupam Datta <\/a>\n <\/h3>\n
\n Principal Research Scientist at Snowflake, Adjunct Professor at Stanford <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n

What is your Agent’s GPA? Toward Trustworthy Data Agents<\/h3>\n

Featured Talk<\/h5>\n
Agents have goals; they plan and act to achieve their goals, often alternating between\u00a0 planning and action steps to refine their plans after reflecting on the results of their actions. Observing that agent failures arise when their goals, plans, and actions are not aligned, we introduce a framework for evaluating and improving an agent\u2019s GPA or Goal-Plan-Action alignment. We create a set of five evaluation metrics \u2013 Goal Fulfillment, Plan Quality, Plan Adherence. Goal Fulfillment measures how well an agent\u2019s actions achieves its goals. Logical Consistency checks that an agent\u2019s actions are consistent with prior actions that it has taken. Execution efficiency checks whether the agent executed in the most efficient way to achieve its goal. Plan Quality checks whether an agent\u2019s plans are aligned with its goals. Finally, Plan Adherence checks if an agent\u2019s actions are aligned with their plan. Our experimental results on two benchmark datasets \u2013 an internal dataset for the Snowflake Intelligence data agent and the public GAIA dataset \u2013 shows that this framework\u00a0 (a) provides a systematic way to cover a broad range of agent failures; (b) exhibits strong agreement between human and LLM judges, ranging from 80% to over 95%; and (c) localizes errors with 85% agreement with human annotations, thus enabling targeted improvement of agent performance.<\/div> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n
\n
\n
\n \n
\n
\n \"Kristian \n <\/div>\n \n \n
\n

\n Max Henrion, PhD <\/h3>\n
\n Founder and CEO, Lumina Decision Systems <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n

How to collaborate with an AI assistant to structure complex decision models with influence diagrams<\/h3>\n

Featured Talk<\/h5>\n
\n
Structuring a complex decision problem with an appropriate frame, creative options, and<\/div>\n
clear tradeoffs are among the most challenging aspects of quality decisions. We are<\/div>\n
starting to learn how generative AI and LLMs can be a useful collaborator in these<\/div>\n
critical tasks. Influence diagrams offer a helpful representation for facilitating the dialog<\/div>\n
between a human decision analyst and AI assistant just as they do between analysts<\/div>\n
and human stakeholders. We demonstrate and discuss examples of how to do this as a<\/div>\n
collaborative process with an AI system to highlight some impressive possibilities as<\/div>\n
well as pitfalls.<\/div>\n<\/div> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n
\n
\n
\n \n
\n
\n \n \"Senior <\/a>\n \n <\/div>\n \n \n
\n

\n \n Dr. Subramanian Iyer <\/a>\n <\/h3>\n
\n Senior Vice President-AI, QXO <\/div>\n \n <\/div>\n \n <\/div>\n\n
\n
\n \n \"Assistant <\/a>\n \n <\/div>\n \n \n
\n

\n \n Dr. Yu Zhang <\/a>\n <\/h3>\n
\n Assistant Professor, UC Santa Cruz <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n \n \"Director <\/a>\n \n <\/div>\n \n \n
\n

\n \n Aravind Govindarajan <\/a>\n <\/h3>\n
\n Director of Data Science, Target Corporation <\/div>\n \n <\/div>\n \n <\/div>\n\n
\n
\n \n \"Pre <\/a>\n \n <\/div>\n \n \n
\n

\n \n Carlos A. Zetina <\/a>\n <\/h3>\n
\n Pre sales Sr. Consultant, FICO <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n

The Future of AI & OR Convergence: How AI and OR Together Redefine Decision-Making<\/h3>\n

Technical Panel<\/h5>\n
\n

As artificial intelligence scales from research labs into real-world systems, operations research provides the rigor and structure that ensures these systems are not only powerful, but also trustworthy, efficient, and sustainable. This interactive panel brings together leading voices at the intersection of AI and OR to explore how these two disciplines complement, challenge, and amplify each other.<\/p>\n

Panelists will discuss the boundaries and synergies of AI and OR, share state-of-the-art applications where optimization unlocks new possibilities for AI, and debate whether OR risks becoming invisible\u2014or more essential than ever\u2014in the age of machine learning. The conversation will also spotlight emerging career pathways, highlight lessons from keynote talks, and address global, ethical, and societal stakes of deploying AI+OR at scale.<\/p>\n

Attendees can expect candid perspectives, debate-provoking contrasts, and practical insights on how to prepare for the future of AI\u2013OR convergence\u2014whether as a researcher, practitioner, or business leader.<\/p>\n<\/div> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n

\n
\n
\n \n
\n
\n \n \"Principal <\/a>\n \n <\/div>\n \n \n
\n

\n \n Keith Dierkx <\/a>\n <\/h3>\n
\n Principal at Princeton Consultants <\/div>\n \n <\/div>\n \n <\/div>\n\n
\n
\n \n \"Sr <\/a>\n \n <\/div>\n \n \n
\n

\n \n Rajat Verma <\/a>\n <\/h3>\n
\n Sr Staff Product Manager, ServiceNow <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n \n \"Lead <\/a>\n \n <\/div>\n \n \n
\n

\n \n Tingting Lin <\/a>\n <\/h3>\n
\n Lead Product Manager, SAP <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n

The Future of Work: How AI\/OR Integration Will Reshape Roles and Skills<\/h3>\n

Business Panel<\/h5>\n
\n

As artificial intelligence and operations research converge, the impact will extend far beyond algorithms and models. It will redefine the very nature of work, skills, and career trajectories. This panel shifts the focus from technical breakthroughs to the human side of the AI & OR synergy: how students, early-career professionals, and industry practitioners can prepare themselves for the next wave of opportunities.<\/p>\n

Bringing together voices from academia, industry, and professional societies, panelists will explore pathways into the AI-OR space, highlight in demand skills and tools, and share candid advice on navigating career transitions. Discussion will spotlight emerging growth industries, strategies for finding mentors and communities, and practical steps to stand out in a competitive market.<\/p>\n

Attendees can expect an engaging conversation that bridges research with real-world careers, covering both the promise and the challenges of this convergence. From actionable recommendations on learning frameworks and building networks to reflections on the future of roles in 2030, this session will equip participants with insights to chart their own journeys at the AI-OR frontier.<\/p>\n<\/div> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n

\n
\n
\n \n
\n
\n \n \"CEO, <\/a>\n \n <\/div>\n \n \n
\n

\n \n Fred Gardi <\/a>\n <\/h3>\n
\n CEO, Hexaly <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n

Hexaly: A super-fast, super-scalable optimization solver<\/h3>\n

Lightning Talk<\/h5>\n
Hexaly is a new type of optimization solver. Its modeling interface is nonlinear and set-oriented. It also supports user-coded functions, enabling seamless integration of simulation with optimization or machine learning with optimization. The Hexaly API unifies modeling concepts from mixed-integer programming, nonlinear programming, and constraint programming. Under the hood, Hexaly combines various exact and heuristic optimization methods: spatial branch-and-bound, simplex methods, interior-point methods, propagation, automatic branch-cut-price, local search, and surrogate modeling.
Hexaly stands out from traditional MILP and CP solvers by delivering super-fast solutions to very large-scale problems like routing, scheduling, sequencing, packing, clustering, matching, assignment, and location. For example, Hexaly provides solutions close to the best-known results in the literature for vehicle routing problems with thousands of points and scheduling problems with millions of tasks, achieving this in just one minute of runtime on a basic computer.<\/div> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n
\n
\n
\n \n
\n
\n \n \"PhD <\/a>\n \n <\/div>\n \n \n
\n

\n \n Yadong (Jeff) Zhang <\/a>\n <\/h3>\n
\n PhD student, Vanderbilt University <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n

Learning Decision Space Structure to Solve Mixed-integer Linear Programs (MILP)<\/h3>\n

Lightning Talk<\/h5>\n
A novel decision space learning framework is proposed for the efficient solution of mixed-integer linear programs (MILPs). The method explicitly models two fundamental aspects of the decision space. First, optimality is represented by a probability distribution defined over the entire feasible region, with higher probability mass assigned to solutions of superior objective value. Second, feasibility of candidate solutions generated through partial variable assignments (i.e., perturbed solutions) is captured by a conditional probability model. The joint probability that a candidate solution is both high-quality and feasible is then evaluated via Bayes\u2019 theorem. To learn this distribution, each MILP instance is encoded as a bipartite graph and a graph neural network is trained to predict the Bayes-based probability of high-quality feasibility, thereby guiding the generation of promising solutions. Extensive experiments on five benchmark MILP classes and real-world instances from the MIPLIB library show that the proposed approach consistently outperforms state-of-the-art solvers (Gurobi and SCIP) as well as a recent machine-learning baseline, achieving up to 23 % higher feasibility rates and up to 24 % improvements in objective value.<\/div> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n
\n
\n
\n \n
\n
\n \n \"Sr. <\/a>\n \n <\/div>\n \n \n
\n

\n \n David Nnamdi <\/a>\n <\/h3>\n
\n Sr. Data Scientist, Intuit <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n

Decomposing the Mix: A Bayesian Distillation Framework for Audience-Level Media Mix Models and Budget Optimization<\/h3>\n

Lightning Talk<\/h5>\n
Marketing Mix Models (MMMs) are analytical staples used to optimize marketing spend, but their traditional assumption of audience homogeneity leads to suboptimal resource allocation. To address this, we present a novel Bayesian distillation framework that translates insights from a trusted, aggregate MMM into a fully specified, audience-level model capable of driving sophisticated optimization. Our approach uses causal insights from audience-level geo-experiments to establish fractional splits that inform unique, audience-specific priors within the new model. A key innovation is a penalization method that ensures consistency between the aggregate and audience-level models, enhancing stability and stakeholder interpretability. This process is made computationally efficient through a JAX-accelerated, two-stage inference strategy combining ADVI and MCMC. The final outputs\u2014robust, audience-specific response curves\u2014are ingested by an optimization solver to allocate marketing spend effectively, enabling the achievement of distinct business goals such as maximizing revenue or acquiring specific high-value audiences.<\/div> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n
\n
\n
\n \n
\n
\n \n \"Applied <\/a>\n \n <\/div>\n \n \n
\n

\n \n Dmitrii Timoshenko <\/a>\n <\/h3>\n
\n Applied Scientist, AWS <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n

Attribution Models: Measuring the Impact of Customer Engagements<\/h3>\n

Lightning Talk<\/h5>\n
In this talk, we will discuss attribution models and their role in solving measurement challenges. We will explore how these models help quantify the impact of different customer engagements on outcomes and overall performance. In addition, we will highlight how incorporating controlled experiments alongside predictive models can improve the accuracy of attribution, providing more reliable insights for decision-making.<\/div> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n
\n
\n
\n \n
\n
\n \n \"Sr. <\/a>\n \n <\/div>\n \n \n
\n

\n \n Dr. Ed Klotz <\/a>\n <\/h3>\n
\n Sr. Mathematical Optimization Specialist <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n

The Current Impact of GPUs and Quantum Computers on Mathematical
Optimization<\/h3>\n

Lightning Talk<\/h5>\n
This talk will survey the promises and challenges of using GPUs and QPUs (Quantum Processing Units) to push forward the state of computational mathematical optimization. GPUs excel at massively parallel execution of simple operations, but these operations do not align neatly with the matrix factorizations central to simplex and barrier methods. Therefore, GPU-friendly algorithmic reformulations or implementations are needed to fully benefit from these processors. Quantum computers, whether annealers or logic-gate models, can initially appear mysterious. Yet, for those familiar with MIP reformulations and the linear algebra underlying optimization, their principles become far more approachable. This perspective also equips practitioners to separate substantive progress from exaggerations and hype. References will be provided for those seeking depth beyond this brief overview.<\/div> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n
\n
\n
\n \n
\n
\n \"Kristian \n <\/div>\n \n \n
\n

\n Dr. Phil Kaminsky <\/h3>\n
\n Senior Principal Research Scientist at Amazon, Professor Emeritus UC Berkeley <\/div>\n \n <\/div>\n \n <\/div>\n <\/div>\n
\n \n
\n
\n

What Executives Really Want from AI & OR<\/h3>\n

Fireside Chat<\/h5>\n
\n

Scientists are continuously advancing the frontiers of AI and Operations Research, but breakthroughs don\u2019t automatically translate into organizational impact. Too often, promising ideas stall because they don\u2019t connect to business priorities\u2014or get swept up in the latest technology hype.<\/p>\n

In this fireside chat, executives will share their perspective on how scientists can shape their work for maximum influence in business settings. They\u2019ll explore:<\/p>\n

    \n
  • Strategies to ensure research delivers measurable business impact.<\/li>\n
  • How to navigate waves of technology hype\u2014from blockchains to LLMs\u2014without losing focus.<\/li>\n
  • Ways scientists can guide executives toward the right opportunities. What the next 3\u20135 years may hold for AI\u2013OR convergence in practice.<\/li>\n<\/ul>\n

    This session is a chance for scientists to hear candidly from business leaders about bridging the gap between research and strategy, positioning themselves as trusted advisors, and future-proofing their careers in an era of rapid change.<\/p>\n<\/div> <\/div>\n<\/div>\n <\/div>\n <\/div>\n <\/div>\n <\/div>\n","protected":false},"excerpt":{"rendered":"

    Speaker Lineup with Talk Titles and Abstracts<\/p>\n","protected":false},"author":1001140,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"_acf_changed":false,"content-type":"","footnotes":""},"class_list":["post-2353","page","type-page","status-publish","hentry","has-post-title","has-post-date","has-post-category","has-post-tag","has-post-comment","has-post-author",""],"acf":[],"yoast_head":"\nSpeaker Lineup with Talk Titles and Abstracts » INFORMS Regional Analytics Conference San Francisco Bay 2025<\/title>\n<meta name=\"description\" content=\"Preview the speaker lineup at the 2025 INFORMS Regional Analytics Conference in San Francisco.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Speaker Lineup with Talk Titles and Abstracts\" \/>\n<meta property=\"og:description\" content=\"Preview the speaker lineup at the 2025 INFORMS Regional Analytics Conference in San Francisco.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/\" \/>\n<meta property=\"og:site_name\" content=\"INFORMS Regional Analytics Conference San Francisco Bay 2025\" \/>\n<meta property=\"article:modified_time\" content=\"2025-09-24T18:53:09+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image7_crop_1.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"300\" \/>\n\t<meta property=\"og:image:height\" content=\"300\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data1\" content=\"1 minute\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/\",\"url\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/\",\"name\":\"Speaker Lineup with Talk Titles and Abstracts » INFORMS Regional Analytics Conference San Francisco Bay 2025\",\"isPartOf\":{\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/#primaryimage\"},\"image\":{\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image7_crop_1.jpg\",\"datePublished\":\"2025-09-24T18:30:35+00:00\",\"dateModified\":\"2025-09-24T18:53:09+00:00\",\"description\":\"Preview the speaker lineup at the 2025 INFORMS Regional Analytics Conference in San Francisco.\",\"breadcrumb\":{\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/#primaryimage\",\"url\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image7_crop_1.jpg\",\"contentUrl\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image7_crop_1.jpg\",\"width\":300,\"height\":300},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Speaker Lineup with Talk Titles and Abstracts\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/#website\",\"url\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/\",\"name\":\"INFORMS Regional Analytics Conference San Francisco Bay 2025\",\"description\":\"Convergence 2025: AI & OR Synergy Summit\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Speaker Lineup with Talk Titles and Abstracts » INFORMS Regional Analytics Conference San Francisco Bay 2025","description":"Preview the speaker lineup at the 2025 INFORMS Regional Analytics Conference in San Francisco.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/","og_locale":"en_US","og_type":"article","og_title":"Speaker Lineup with Talk Titles and Abstracts","og_description":"Preview the speaker lineup at the 2025 INFORMS Regional Analytics Conference in San Francisco.","og_url":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/","og_site_name":"INFORMS Regional Analytics Conference San Francisco Bay 2025","article_modified_time":"2025-09-24T18:53:09+00:00","og_image":[{"width":300,"height":300,"url":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image7_crop_1.jpg","type":"image\/jpeg"}],"twitter_card":"summary_large_image","twitter_misc":{"Est. reading time":"1 minute"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/","url":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/","name":"Speaker Lineup with Talk Titles and Abstracts » INFORMS Regional Analytics Conference San Francisco Bay 2025","isPartOf":{"@id":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/#website"},"primaryImageOfPage":{"@id":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/#primaryimage"},"image":{"@id":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/#primaryimage"},"thumbnailUrl":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image7_crop_1.jpg","datePublished":"2025-09-24T18:30:35+00:00","dateModified":"2025-09-24T18:53:09+00:00","description":"Preview the speaker lineup at the 2025 INFORMS Regional Analytics Conference in San Francisco.","breadcrumb":{"@id":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/#primaryimage","url":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image7_crop_1.jpg","contentUrl":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image7_crop_1.jpg","width":300,"height":300},{"@type":"BreadcrumbList","@id":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/lineup\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/"},{"@type":"ListItem","position":2,"name":"Speaker Lineup with Talk Titles and Abstracts"}]},{"@type":"WebSite","@id":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/#website","url":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/","name":"INFORMS Regional Analytics Conference San Francisco Bay 2025","description":"Convergence 2025: AI & OR Synergy Summit","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"}]}},"builder_content":"<h1 style=\"text-align: left;\">Speaker Lineup with Talk Titles and Abstracts<\/h1>\n<a href=\"https:\/\/www.linkedin.com\/in\/anima-anandkumar\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image7_crop_1.jpg\" title=\"Dr. Anima Anandkumar\" alt=\"Kristian Hammond Keynote at 2024 INFORMS Regional Analytics Chicago\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image7_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image7_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/anima-anandkumar\/\" rel=\"noopener\" target=\"_blank\"> Dr. Anima Anandkumar <\/a> <\/h3> Bren Professor at CalTech\n<h3>Neural Operators for AI+Science: Pushing the Frontiers of Scientific Discovery<\/h3> <h5>Keynote<\/h5> The main bottleneck in doing scientific research is the need for physical experiments in many areas. This means risky ideas are often discarded and the hypothesis space is traditionally restricted to regions of prior success. AI is disrupting this status quo by enabling physically-valid digital twins that reduce or even completely remove the need for physical experiments. AI models are orders of magnitude faster than traditional simulations, and often more accurate, since they can directly adapt to experimental and observational data. Since AI models are differentiable, they can be directly used for inverse design, enabling exploration and design optimization subject to diverse constraints, that was not possible before. Neural Operators enable multiscale and physics-informed learning for achieving high fidelity and training data efficiency in many areas. They have been successfully applied in weather and climate modeling, plasma evolution in nuclear fusion, designing novel medical devices and enabling autonomous flights under turbulence.\n<a href=\"https:\/\/www.linkedin.com\/in\/anupamdatta\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image6_crop_1.jpg\" title=\"Dr. Anupam Datta\" alt=\"Kristian Hammond Keynote at 2024 INFORMS Regional Analytics Chicago\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image6_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image6_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/anupamdatta\/\" rel=\"noopener\" target=\"_blank\"> Dr. Anupam Datta <\/a> <\/h3> Principal Research Scientist at Snowflake, Adjunct Professor at Stanford\n<h3>What is your Agent's GPA? Toward Trustworthy Data Agents<\/h3> <h5>Featured Talk<\/h5> Agents have goals; they plan and act to achieve their goals, often alternating between\u00a0 planning and action steps to refine their plans after reflecting on the results of their actions. Observing that agent failures arise when their goals, plans, and actions are not aligned, we introduce a framework for evaluating and improving an agent\u2019s GPA or Goal-Plan-Action alignment. We create a set of five evaluation metrics \u2013 Goal Fulfillment, Plan Quality, Plan Adherence. Goal Fulfillment measures how well an agent\u2019s actions achieves its goals. Logical Consistency checks that an agent\u2019s actions are consistent with prior actions that it has taken. Execution efficiency checks whether the agent executed in the most efficient way to achieve its goal. Plan Quality checks whether an agent\u2019s plans are aligned with its goals. Finally, Plan Adherence checks if an agent\u2019s actions are aligned with their plan. Our experimental results on two benchmark datasets \u2013 an internal dataset for the Snowflake Intelligence data agent and the public GAIA dataset \u2013 shows that this framework\u00a0 (a) provides a systematic way to cover a broad range of agent failures; (b) exhibits strong agreement between human and LLM judges, ranging from 80% to over 95%; and (c) localizes errors with 85% agreement with human annotations, thus enabling targeted improvement of agent performance.\n<img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Max_Henrion.png\" title=\"Max Henrion, PhD\" alt=\"Kristian Hammond Keynote at 2024 INFORMS Regional Analytics Chicago\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Max_Henrion.png 200w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Max_Henrion-150x150.png 150w\" sizes=\"(max-width: 200px) 100vw, 200px\" \/> <h3> Max Henrion, PhD <\/h3> Founder and CEO, Lumina Decision Systems\n<h3>How to collaborate with an AI assistant to structure complex decision models with influence diagrams<\/h3> <h5>Featured Talk<\/h5>\nStructuring a complex decision problem with an appropriate frame, creative options, and clear tradeoffs are among the most challenging aspects of quality decisions. We are starting to learn how generative AI and LLMs can be a useful collaborator in these critical tasks. Influence diagrams offer a helpful representation for facilitating the dialog between a human decision analyst and AI assistant just as they do between analysts and human stakeholders. We demonstrate and discuss examples of how to do this as a collaborative process with an AI system to highlight some impressive possibilities as well as pitfalls.\n<a href=\"https:\/\/www.linkedin.com\/in\/skiyer\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image3_crop_1.jpg\" title=\"Dr. Subramanian Iyer\" alt=\"Senior Vice President-AI, QXO\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image3_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image3_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/skiyer\/\" rel=\"noopener\" target=\"_blank\"> Dr. Subramanian Iyer <\/a> <\/h3> Senior Vice President-AI, QXO\n<a href=\"https:\/\/www.linkedin.com\/in\/yu-zhang-8b276320\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image5_crop_1.jpg\" title=\"Dr. Yu Zhang\" alt=\"Assistant Professor, UC Santa Cruz\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image5_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image5_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/yu-zhang-8b276320\/\" rel=\"noopener\" target=\"_blank\"> Dr. Yu Zhang <\/a> <\/h3> Assistant Professor, UC Santa Cruz\n<a href=\"https:\/\/www.linkedin.com\/in\/aravind-govindarajan\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image11_crop_1.jpg\" title=\"Aravind Govindarajan\" alt=\"Director of Data Science, Target Corporation\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image11_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image11_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/aravind-govindarajan\/\" rel=\"noopener\" target=\"_blank\"> Aravind Govindarajan <\/a> <\/h3> Director of Data Science, Target Corporation\n<a href=\"https:\/\/www.linkedin.com\/in\/cazetina\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Carlos_Zetina_crop_2.jpg\" title=\"Carlos A. Zetina\" alt=\"Pre sales Sr. Consultant, FICO\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Carlos_Zetina_crop_2.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Carlos_Zetina_crop_2-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/cazetina\/\" rel=\"noopener\" target=\"_blank\"> Carlos A. Zetina <\/a> <\/h3> Pre sales Sr. Consultant, FICO\n<h3>The Future of AI & OR Convergence: How AI and OR Together Redefine Decision-Making<\/h3> <h5>Technical Panel<\/h5>\n<p>As artificial intelligence scales from research labs into real-world systems, operations research provides the rigor and structure that ensures these systems are not only powerful, but also trustworthy, efficient, and sustainable. This interactive panel brings together leading voices at the intersection of AI and OR to explore how these two disciplines complement, challenge, and amplify each other.<\/p> <p>Panelists will discuss the boundaries and synergies of AI and OR, share state-of-the-art applications where optimization unlocks new possibilities for AI, and debate whether OR risks becoming invisible\u2014or more essential than ever\u2014in the age of machine learning. The conversation will also spotlight emerging career pathways, highlight lessons from keynote talks, and address global, ethical, and societal stakes of deploying AI+OR at scale.<\/p> <p>Attendees can expect candid perspectives, debate-provoking contrasts, and practical insights on how to prepare for the future of AI\u2013OR convergence\u2014whether as a researcher, practitioner, or business leader.<\/p>\n<a href=\"https:\/\/www.linkedin.com\/in\/keithdierkx\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Keith_Dierkx_crop_1.jpg\" title=\"Keith Dierkx\" alt=\"Principal at Princeton Consultants\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Keith_Dierkx_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Keith_Dierkx_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/keithdierkx\/\" rel=\"noopener\" target=\"_blank\"> Keith Dierkx <\/a> <\/h3> Principal at Princeton Consultants\n<a href=\"https:\/\/www.linkedin.com\/in\/rajatverma22\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image8_crop_1.jpg\" title=\"Rajat Verma\" alt=\"Sr Staff Product Manager, ServiceNow\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image8_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image8_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/rajatverma22\/\" rel=\"noopener\" target=\"_blank\"> Rajat Verma <\/a> <\/h3> Sr Staff Product Manager, ServiceNow\n<a href=\"https:\/\/www.linkedin.com\/in\/tingting-lin\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Tingting_Lin_crop_3.jpg\" title=\"Tingting Lin\" alt=\"Lead Product Manager, SAP\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Tingting_Lin_crop_3.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/09\/Tingting_Lin_crop_3-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/tingting-lin\/\" rel=\"noopener\" target=\"_blank\"> Tingting Lin <\/a> <\/h3> Lead Product Manager, SAP\n<h3>The Future of Work: How AI\/OR Integration Will Reshape Roles and Skills<\/h3> <h5>Business Panel<\/h5>\n<p>As artificial intelligence and operations research converge, the impact will extend far beyond algorithms and models. It will redefine the very nature of work, skills, and career trajectories. This panel shifts the focus from technical breakthroughs to the human side of the AI & OR synergy: how students, early-career professionals, and industry practitioners can prepare themselves for the next wave of opportunities.<\/p> <p>Bringing together voices from academia, industry, and professional societies, panelists will explore pathways into the AI-OR space, highlight in demand skills and tools, and share candid advice on navigating career transitions. Discussion will spotlight emerging growth industries, strategies for finding mentors and communities, and practical steps to stand out in a competitive market.<\/p> <p>Attendees can expect an engaging conversation that bridges research with real-world careers, covering both the promise and the challenges of this convergence. From actionable recommendations on learning frameworks and building networks to reflections on the future of roles in 2030, this session will equip participants with insights to chart their own journeys at the AI-OR frontier.<\/p>\n<a href=\"https:\/\/www.linkedin.com\/in\/fredericgardi\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image1_crop_1.jpg\" title=\"Fred Gardi\" alt=\"CEO, Hexaly\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image1_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image1_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/fredericgardi\/\" rel=\"noopener\" target=\"_blank\"> Fred Gardi <\/a> <\/h3> CEO, Hexaly\n<h3>Hexaly: A super-fast, super-scalable optimization solver<\/h3> <h5>Lightning Talk<\/h5> Hexaly is a new type of optimization solver. Its modeling interface is nonlinear and set-oriented. It also supports user-coded functions, enabling seamless integration of simulation with optimization or machine learning with optimization. The Hexaly API unifies modeling concepts from mixed-integer programming, nonlinear programming, and constraint programming. Under the hood, Hexaly combines various exact and heuristic optimization methods: spatial branch-and-bound, simplex methods, interior-point methods, propagation, automatic branch-cut-price, local search, and surrogate modeling.<br>Hexaly stands out from traditional MILP and CP solvers by delivering super-fast solutions to very large-scale problems like routing, scheduling, sequencing, packing, clustering, matching, assignment, and location. For example, Hexaly provides solutions close to the best-known results in the literature for vehicle routing problems with thousands of points and scheduling problems with millions of tasks, achieving this in just one minute of runtime on a basic computer.\n<a href=\"https:\/\/www.linkedin.com\/in\/fredericgardi\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image4_crop_1.jpg\" title=\"Yadong (Jeff) Zhang\" alt=\"PhD student, Vanderbilt University\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image4_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image4_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/fredericgardi\/\" rel=\"noopener\" target=\"_blank\"> Yadong (Jeff) Zhang <\/a> <\/h3> PhD student, Vanderbilt University\n<h3>Learning Decision Space Structure to Solve Mixed-integer Linear Programs (MILP)<\/h3> <h5>Lightning Talk<\/h5> A novel decision space learning framework is proposed for the efficient solution of mixed-integer linear programs (MILPs). The method explicitly models two fundamental aspects of the decision space. First, optimality is represented by a probability distribution defined over the entire feasible region, with higher probability mass assigned to solutions of superior objective value. Second, feasibility of candidate solutions generated through partial variable assignments (i.e., perturbed solutions) is captured by a conditional probability model. The joint probability that a candidate solution is both high-quality and feasible is then evaluated via Bayes\u2019 theorem. To learn this distribution, each MILP instance is encoded as a bipartite graph and a graph neural network is trained to predict the Bayes-based probability of high-quality feasibility, thereby guiding the generation of promising solutions. Extensive experiments on five benchmark MILP classes and real-world instances from the MIPLIB library show that the proposed approach consistently outperforms state-of-the-art solvers (Gurobi and SCIP) as well as a recent machine-learning baseline, achieving up to 23 % higher feasibility rates and up to 24 % improvements in objective value.\n<a href=\"https:\/\/www.linkedin.com\/in\/david-nnamdi\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image2_crop_1.jpg\" title=\"David Nnamdi\" alt=\"Sr. Data Scientist, Intuit\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image2_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image2_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/david-nnamdi\/\" rel=\"noopener\" target=\"_blank\"> David Nnamdi <\/a> <\/h3> Sr. Data Scientist, Intuit\n<h3>Decomposing the Mix: A Bayesian Distillation Framework for Audience-Level Media Mix Models and Budget Optimization<\/h3> <h5>Lightning Talk<\/h5> Marketing Mix Models (MMMs) are analytical staples used to optimize marketing spend, but their traditional assumption of audience homogeneity leads to suboptimal resource allocation. To address this, we present a novel Bayesian distillation framework that translates insights from a trusted, aggregate MMM into a fully specified, audience-level model capable of driving sophisticated optimization. Our approach uses causal insights from audience-level geo-experiments to establish fractional splits that inform unique, audience-specific priors within the new model. A key innovation is a penalization method that ensures consistency between the aggregate and audience-level models, enhancing stability and stakeholder interpretability. This process is made computationally efficient through a JAX-accelerated, two-stage inference strategy combining ADVI and MCMC. The final outputs\u2014robust, audience-specific response curves\u2014are ingested by an optimization solver to allocate marketing spend effectively, enabling the achievement of distinct business goals such as maximizing revenue or acquiring specific high-value audiences.\n<a href=\"https:\/\/www.linkedin.com\/in\/dm-timoshenko\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image10_crop_1.jpg\" title=\"Dmitrii Timoshenko\" alt=\"Applied Scientist, AWS\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image10_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image10_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/dm-timoshenko\/\" rel=\"noopener\" target=\"_blank\"> Dmitrii Timoshenko <\/a> <\/h3> Applied Scientist, AWS\n<h3>Attribution Models: Measuring the Impact of Customer Engagements<\/h3> <h5>Lightning Talk<\/h5> In this talk, we will discuss attribution models and their role in solving measurement challenges. We will explore how these models help quantify the impact of different customer engagements on outcomes and overall performance. In addition, we will highlight how incorporating controlled experiments alongside predictive models can improve the accuracy of attribution, providing more reliable insights for decision-making.\n<a href=\"https:\/\/www.linkedin.com\/in\/ed-klotz-2395471\/\" rel=\"noopener\" target=\"_blank\"> <img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image9_crop_1.jpg\" title=\"Dr. Ed Klotz\" alt=\"Sr. Mathematical Optimization Specialist\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image9_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/image9_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <\/a> <h3> <a href=\"https:\/\/www.linkedin.com\/in\/ed-klotz-2395471\/\" rel=\"noopener\" target=\"_blank\"> Dr. Ed Klotz <\/a> <\/h3> Sr. Mathematical Optimization Specialist\n<h3>The Current Impact of GPUs and Quantum Computers on Mathematical<br>Optimization<\/h3> <h5>Lightning Talk<\/h5> This talk will survey the promises and challenges of using GPUs and QPUs (Quantum Processing Units) to push forward the state of computational mathematical optimization. GPUs excel at massively parallel execution of simple operations, but these operations do not align neatly with the matrix factorizations central to simplex and barrier methods. Therefore, GPU-friendly algorithmic reformulations or implementations are needed to fully benefit from these processors. Quantum computers, whether annealers or logic-gate models, can initially appear mysterious. Yet, for those familiar with MIP reformulations and the linear algebra underlying optimization, their principles become far more approachable. This perspective also equips practitioners to separate substantive progress from exaggerations and hype. References will be provided for those seeking depth beyond this brief overview.\n<img src=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/Phil_Kaminsky_crop_1.jpg\" title=\"Dr. Phil Kaminsky\" alt=\"Kristian Hammond Keynote at 2024 INFORMS Regional Analytics Chicago\" srcset=\"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/Phil_Kaminsky_crop_1.jpg 300w, https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/files\/2025\/08\/Phil_Kaminsky_crop_1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/> <h3> Dr. Phil Kaminsky <\/h3> Senior Principal Research Scientist at Amazon, Professor Emeritus UC Berkeley\n<h3>What Executives Really Want from AI & OR<\/h3> <h5>Fireside Chat<\/h5>\n<p>Scientists are continuously advancing the frontiers of AI and Operations Research, but breakthroughs don\u2019t automatically translate into organizational impact. Too often, promising ideas stall because they don\u2019t connect to business priorities\u2014or get swept up in the latest technology hype.<\/p> <p>In this fireside chat, executives will share their perspective on how scientists can shape their work for maximum influence in business settings. They\u2019ll explore:<\/p> <ul> <li>Strategies to ensure research delivers measurable business impact.<\/li> <li>How to navigate waves of technology hype\u2014from blockchains to LLMs\u2014without losing focus.<\/li> <li>Ways scientists can guide executives toward the right opportunities. What the next 3\u20135 years may hold for AI\u2013OR convergence in practice.<\/li> <\/ul> <p>This session is a chance for scientists to hear candidly from business leaders about bridging the gap between research and strategy, positioning themselves as trusted advisors, and future-proofing their careers in an era of rapid change.<\/p>","_links":{"self":[{"href":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/wp-json\/wp\/v2\/pages\/2353","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/wp-json\/wp\/v2\/users\/1001140"}],"replies":[{"embeddable":true,"href":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/wp-json\/wp\/v2\/comments?post=2353"}],"version-history":[{"count":24,"href":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/wp-json\/wp\/v2\/pages\/2353\/revisions"}],"predecessor-version":[{"id":2377,"href":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/wp-json\/wp\/v2\/pages\/2353\/revisions\/2377"}],"wp:attachment":[{"href":"https:\/\/meetings.informs.org\/wordpress\/regional-analytics-sanfrancisco2025\/wp-json\/wp\/v2\/media?parent=2353"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}