Sign in

Connor Lawless

@lawlessopt.bsky.social
454 followers 304 following 27 posts

Stanford MS&E Postdoc | Human-Centered AI & OR Prev: @CornellORIE @MSFTResearch, @IBMResearch, @uoftmie 🌈

PostsRepliesMedia
Reposted by Connor Lawless
Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 30/09/2026
In our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. www.nature.com/articles/s41...
nature.com
Scalable decision-making for games of imperfect information - Nature
Ataraxos, an AI for the board wargame Stratego, establishes a design pattern for reinforcement learning and search that is effective under large amounts of hidden information, a longstanding desiderat...
1225960
Reposted by Connor Lawless
Ellen Vitercik @ellen-v.bsky.social · 26/08/2026
LLMs can make optimization more accessible, but they must not sacrifice the guarantees that make solvers trustworthy. This was the focus of my part of our @ijcai.org tutorial, Large Language Models for Optimization, w/ Fei Liu & @lawlessopt.bsky.social. Materials: feiliu36.github.io/llm_opt_tuto...
052
Reposted by Connor Lawless
Sophie Huiberts @sophie.huiberts.me · 30/05/2026
Krunal Patel makes a brilliant video series about the internals of MIP solvers. One of the best resources for learning
youtu.be
Cut Pool Management in SCIP, HiGHS, and CP-SAT
YouTube video by Krunal Patel
0112
Reposted by Connor Lawless
Ellen Vitercik @ellen-v.bsky.social · 20/01/2026
@lawlessopt.bsky.social and I are excited to present our #AAAI2026 tutorial on “LLMs for Optimization: Modeling, Solving, and Validating with Generative AI.” When: Tuesday, Jan 20, 2026, 8:30am–12:30pm SGT Where: Garnet 216 (Singapore EXPO) (Connor’s intro slides are shown here.) CC @aaai.org
181
Connor Lawless @lawlessopt.bsky.social · 18/11/2025
Thank you!!
000
Connor Lawless @lawlessopt.bsky.social · 16/11/2025
It's been an absolute pleasure working with Ellen, Madeleine, and their amazing PhD students for the past year on making optimization more accessible with generative AI! I am on the job market this year - check out my website (conlaw.github.io) for more details on what I've been up to.
conlaw.github.io
Connor Lawless
173
Reposted by Connor Lawless
CPAIOR @cpaior.bsky.social · 12/11/2025
In our final session for the day, we're focused on a hot topic: machine learning and mixed integer programming. Connor Lawless will start the session and tells us how to use LLMs for cold-start cutting plane separator configuration. doi.org/10.1007/978-...
141
Reposted by Connor Lawless
Cathy Wu @cathywu.bsky.social · 12/08/2025
🔥 New workshop at @neuripsconf.bsky.social! DiffCoALG bridges the gap between classic algorithms & differentiable learning. Think: LLM reasoning, routing, SAT, MIP — neurally optimized. 📌 Submit by Aug 22! 🤖🧠 🔗 sites.google.com/view/diffcoa... #NeurIPS2025
sites.google.com
DiffCoALG@NeurIPS 2025
About this Workshop Combinatorial algorithms are fundamental across a wide range of domains, owing to their ability to model optimization and decision-making tasks under complex constraints. These alg...
1104
Connor Lawless @lawlessopt.bsky.social · 16/07/2025
📕: EquivaMap: Leveraging LLMs for Automatice Equivalence Checking of Optimization formulations (Joint work with @ellen-v.bsky.social @hzhai.bsky.social and @leqiliu.bsky.social ) 🔗: arxiv.org/abs/2502.14760
arxiv.org
EquivaMap: Leveraging LLMs for Automatic Equivalence Checking of Optimization Formulations
A fundamental problem in combinatorial optimization is identifying equivalent formulations. Despite the growing need for automated equivalence checks -- driven, for example, by optimization copilots, ...
030
Connor Lawless @lawlessopt.bsky.social · 16/07/2025
There's been a lot of work using LLMs to formulate MILPs, but how do we know that the formulations are correct? Come chat with Haotian at poster W-515 to learn about our work on automatic equivalence checking for optimization models!
160
Connor Lawless @lawlessopt.bsky.social · 14/07/2025
Our empirical results highlight that existing pointwise approaches for recourse can fail to catch potential fixed predictions, whereas our approach (provably) succeeds!
010
Connor Lawless @lawlessopt.bsky.social · 14/07/2025
We model the problem as a mixed-integer quadratically constrained program that runs in seconds on real-world datasets.
110
Connor Lawless @lawlessopt.bsky.social · 14/07/2025
This paradigm lets us spot fixed predictions before deploying a model, lets us audit public models for recourse (even if we don't have any available data!), and gives interpretable summaries of regions with fixed predictions to help with debugging.
110
Connor Lawless @lawlessopt.bsky.social · 14/07/2025
In this paper, we introduce a new paradigm for algorithmic recourse that aims to certify recourse over an entire region of the feature space!
110
Connor Lawless @lawlessopt.bsky.social · 14/07/2025
Existing approaches to algorithmic recourse focus on verifying recourse on an individual-by-individual basis, which can cause model developers to miss potential fixed predictions, requires a lot of data, and makes it difficult to debug recourse issues!
110
Connor Lawless @lawlessopt.bsky.social · 14/07/2025
Machine learning models can assign fixed predictions that preclude individuals from changing their outcome. Think credit applicants that can never get a loan approved, or young patients that can never get an organ transplant - no matter how sick they are!
111
Connor Lawless @lawlessopt.bsky.social · 14/07/2025
Excited to be chatting about our new paper "Understanding Fixed Predictions via Confined Regions" (joint work with @berkustun.bsky.social, Lily Weng, and Madeleine Udell) at #ICML2025! 🕐 Wed 16 Jul 4:30 p.m. PDT — 7 p.m. PDT 📍East Exhibition Hall A-B #E-1104 🔗 arxiv.org/abs/2502.16380
arxiv.org
Understanding Fixed Predictions via Confined Regions
Machine learning models can assign fixed predictions that preclude individuals from changing their outcome. Existing approaches to audit fixed predictions do so on a pointwise basis, which requires ac...
153
Reposted by Connor Lawless
Yu He @dransyhe.bsky.social · 13/07/2025
Our ✨spotlight paper✨ "Primal-Dual Neural Algorithmic Reasoning" is coming to #ICML2025! We bring Neural Algorithmic Reasoning (NAR) to the NP-hard frontier 💥 🗓 Poster session: Tuesday 11:00–13:30 📍 East Exhibition Hall A-B, # E-3003 🔗 openreview.net/pdf?id=iBpkz... 🧵
162
Connor Lawless @lawlessopt.bsky.social · 25/03/2025
This is my first time at an HCI conference - come say hi if you're around!
010
Connor Lawless @lawlessopt.bsky.social · 25/03/2025
In addition to a bunch of quantitative experiments, we ran a user study with a prototype system to inform design recommendations for future interactive optimization systems. Check out the paper for more details!
110
Connor Lawless @lawlessopt.bsky.social · 25/03/2025
We built a hybrid LLM and CP system that uses LLMs to translate user requests in chat into operations on an underlying CP optimization model to schedule a new meeting. This gets the best of both worlds - the flexibility of LLMs with the decision making power of optimization!
120
Connor Lawless @lawlessopt.bsky.social · 25/03/2025
Building optimization models in practice involves a ton of back and forth between optimization and domain experts to understand a decision making problem. Can we enable domain experts to craft their own optimization models instead? We study this through the lens of scheduling.
110
Connor Lawless @lawlessopt.bsky.social · 25/03/2025
Excited to be chatting about our ACM TIIS paper at IUI today: "I Want it That Way": Enabling Interactive Decision Support via Large Language Models and Constraint Programming 🔗: arxiv.org/abs/2312.06908
arxiv.org
"I Want It That Way": Enabling Interactive Decision Support Using Large Language Models and Constraint Programming
A critical factor in the success of decision support systems is the accurate modeling of user preferences. Psychology research has demonstrated that users often develop their preferences during the el...
130
Connor Lawless @lawlessopt.bsky.social · 16/03/2025
In case youre wondering why this thread looks suspiciously like a bunch of screenshots from a presentation... I'll be chatting about this project at the INFORMs Computing Society Conference in the debate room at 3. Come say hi!
010
Connor Lawless @lawlessopt.bsky.social · 16/03/2025
More broadly, this is a first step towards a new paradigm where we can exploit natural language information to do better algorithm configuration and design! There's tons of exciting open problems towards this goal (reach out if you're interested!).
010
Connor Lawless @lawlessopt.bsky.social · 16/03/2025
Surprisingly, we can get high performing configurations from our framework - outperforming solver defaults on a number of real world problems, without solving a single MILP!
100
Connor Lawless @lawlessopt.bsky.social · 16/03/2025
We introduce a LLM based framework with some algorithmic bells and whistles (ensembling, solver specific context...) to capitalize on LLM strengths while addressing these challenges.
100
Connor Lawless @lawlessopt.bsky.social · 16/03/2025
Unfortunately, LLMs aren't a natural fit for configuration. Parameters are problem specific, LLMs have stochastic outputs, and frankly - it's a tough problem!
100
Connor Lawless @lawlessopt.bsky.social · 16/03/2025
Can we get better problem-specific solver configurations without the big computational price tag? In this paper we show that we can thanks to Large Language Models! Why LLMs? They can identify useful optimization structure and have a lot of built in math programming knowledge!
111
Connor Lawless @lawlessopt.bsky.social · 16/03/2025
MILP solvers ship with a ton of parameters that can have a massive impact on solver performance (over 70% for separator configuration alone!), but are notoriously difficult to set. Existing approaches for algorithm configuration require solving a ton of MILPs leading to days of compute.
210
Connor Lawless @lawlessopt.bsky.social · 16/03/2025
Super excited about this new work with Yingxi Li, Anders Wikun, @ellen-v.bsky.social, and Madeleine Udell forthcoming at CPAIOR2025: LLMs for Cold-Start Cutting Plane Separator Configuration 🔗: arxiv.org/abs/2412.12038
arxiv.org
LLMs for Cold-Start Cutting Plane Separator Configuration
Mixed integer linear programming (MILP) solvers ship with a staggering number of parameters that are challenging to select a priori for all but expert optimization users, but can have an outsized impa...
1115
Reposted by Connor Lawless
Karen Hao @karenhao.bsky.social · 21/02/2025
For decades, the US government has painstakingly kept American science #1 globally—and every facet of American life has improved because of it. The internet? Flu shot? Ozempic? All grew out of federally-funded research. Now all that's being dismantled. 1/ www.technologyreview.com/2025/02/21/1...
technologyreview.com
The foundations of America’s prosperity are being dismantled
Federal scientists warn that Americans could feel the effects of the new administration's devastating cuts for decades to come
7829981496
Reposted by Connor Lawless
angela zhou @angelamczhou.bsky.social · 12/02/2025
"Not a step back" Possibly --- even a step _forward_? /s
131
Reposted by Connor Lawless
Thiago Serra @thserra.bsky.social · 05/02/2025
Summer 2025 schools on algorithms, machine learning, mathematics, optimization, and other relevant topics in operations research #orms
thiagoserra.com
Summer 2025 schools on algorithms, machine learning, mathematics, optimization, and other relevant topics in operations research
After a 5-year hiatus and finding out that this was helpful to many people, I am reviving the list of schools. The purpose of organizing this is to help graduate students find a summer school to ga…
073
Connor Lawless @lawlessopt.bsky.social · 03/02/2025
The goal of the study is to inform the future design of automated modelling tools (think GitHub copilot for optimizers) - (in addition to my boundless gratitude) your feedback has the chance to shape years of research! Please consider participating!
010
Connor Lawless @lawlessopt.bsky.social · 03/02/2025
Have you developed and/or implemented an optimization model to solve a real-world use case? We want to hear from you! We're extending our study on workflows in optimization modelling for a couple last interviews: tinyurl.com/3ejdr7su
142
Reposted by Connor Lawless
Dolores Romero Morales @doloresromerom.bsky.social · 31/01/2025
Yesterday, I gave a @gurobioptimization.bsky.social webinar where I explored how to make #AI and #ML both #transparent and #fair without compromising on #accuracy. The recording of the talk is at www.brighttalk.com/webcast/1918.... Happy to hear about your feedback!
brighttalk.com
On a (Short) Optimization Tour Through Transparent and Fair ML with Prof. Dolores Morales
There is a common consensus that state-of-the-art Artificial Intelligence (AI) and Machine Learning (ML) algorithms are powerful in terms of their accuracy, but they are also perceived as opaque not b...
0102
Reposted by Connor Lawless
Thiago Serra @thserra.bsky.social · 09/01/2025
My department is looking for a postdoc. We have great (and nice!) folks working on theory and applications in data science, machine learning, mathematical optimization, and transportation. The university also has a lot going on in heathcare. See more below: jobs.uiowa.edu/postdoc/view...
jobs.uiowa.edu
Postdoctoral Requisition Details - Jobs@UIOWA: Search and Apply for Jobs at The University of IowaUniversity of IowaUniversity of Iowa
Jobs@UIOWA: The official place to search and apply for jobs at The University of Iowa.
094
Connor Lawless @lawlessopt.bsky.social · 27/11/2024
Awesome blog! You should check out some of the work coming out of academia on LLMs for optimization modelling: optimus-solver.com/login and paper: www.arxiv.org/abs/2407.19633
optimus-solver.com
Optimus
Web site created using create-react-app
061