=========================== WELCOME TO THE TERMINAL =========================== Hello! I'm Aristides, a Ph.D. student in machine learning, and welcome to my website! Click one of the commands below to get started, or check out my contact info: - Email: aristides.milios [at] mila.quebec - LinkedIn: linkedin.com/in/aristidesmilios - GitHub: github.com/amilios - Google Scholar: Google Scholar Profile - Twitter: @amilios

============= BIOGRAPHY ============= I'm a machine learning researcher focusing on LLMs and hallucination reduction/uncertainty prediction/learning to abstain. I co-founded a startup called Bitness.io in high school, created a program for kids called Hoist Halifax to promote entrepreneurship, and led design thinking workshops at Dalhousie as part of Stanford's d.school's "University Innovation Fellows" program. I've done research internships at MarineTraffic (Athens) and Know-Center (Graz, Austria), as well as MITACS internships at ServiceNow Research (former ElementAI) in Montreal, Canada. I have a B.Comp.Sci from Dalhousie (First Class Honours), an M.Sc. from McGill/Mila under Siva Reddy and Dzmitry Bahdanau, and am currently working on a Ph.D. at Université de Montréal under Chris Pal.

================== RESEARCH FOCUS ================== My current research centers on improving the reliability and trustworthiness of large language models, particularly focusing on: - Hallucination detection and reduction techniques - Uncertainty quantification in neural language models - Teaching LLMs to abstain when they are uncertain enough

================ PUBLICATIONS ================ - LLMs can learn self-restraint through iterative self-reflection (ICLR 2024 Workshop) - In-Context Learning for Text Classification with Many Labels (GenBench Workshop at EMNLP 2023) - An Analysis of Social Biases Present in BERT Variants Across Multiple Languages (TSRML Workshop at NeurIPS 2022) - Survey of Generative Methods for Social Media Analysis (ArXiv report) - Monitoring Marine Protected Areas using Data Fusion and AI Techniques (NATO CMRE 2019) - Automatic Fusion of Satellite Imagery and AIS data for Vessel Detection (FUSION 2019)

============ TIMELINE ============ Summer 2018, Winter 2019: MarineTraffic internships (Athens) Fall 2019: Know-Center internship (Austria) August 2020: graduated B.Comp.Sci First Class Honours, Dalhousie September 2021: started M.Sc. at McGill/Mila, co-supervision Siva Reddy/Dzmitry Bahdanau Spring 2024: graduated M.Sc. September 2024: started Ph.D. at Université de Montréal March 2025: started MITACS Accelerate internship at ServiceNow Research

=================== WORK EXPERIENCE =================== - NACO Canada - Contractor (Jan 2021 - Aug 2023) -> Designed automation system for yearly investment reports - Dalhousie University - Student Researcher (Winter 2021) -> Healthcare time series clustering with Dr. Luis Torgo -> NLP societal risks research with Dr. Stan Matwin Internships: - Know-Center GmbH, Graz, Austria (Fall 2019) -> Active Learning for Recommender Systems, RNNs with Attention - MarineTraffic Inc., Athens, Greece (Summer 2018, Winter 2019) -> AIS data fusion with satellite imagery for maritime awareness

======================= TEACHING EXPERIENCE ======================= Fall 2022, Spring 2024: Teaching Assistant at McGill University Montreal, Quebec, Canada - Teaching Assistant for Dr. Siva Reddy for course COMP 599 "Natural Language Understanding with Deep Learning/Computational Semantics" - Designed an assignment from scratch on RNN/LSTM/BiLSTM networks - students implemented cells using basic PyTorch building blocks Fall and Winter 2021: Teaching Assistant at Dalhousie University Halifax, Nova Scotia, Canada - CSCI 3162 Digital Media, with Dr. Dirk Arnold (Winter 2021) - CSCI 3141 Foundations of Data Science using R, with Dr. Luis Torgo (Fall 2020) - CSCI 2170 Server-side Scripting, with Dr. Bonnie MacKay (Fall 2020) - CSCI 2141 Intro to Database Systems, with Dr. Khurram Aziz (Fall 2020)

========================= AWARDS & SCHOLARSHIPS ========================= Major Scholarships: - IVADO M.Sc. Excellence Scholarship - $40,000 CAD - NSERC CGS-M Grant - $17,500 CAD - Dalhousie Schulich Scholarship - $41,900 CAD Other Awards: - McGill Lorne Trottier Graduate Fellowship - $5,000 CAD - MITACS Globalink Grant - $6,000 CAD - Dalhousie Bruce and Dorothy Rossetti Scholarship Silver - $1,750 CAD

============== LEADERSHIP ============== Academic Leadership: - Laboratory Representative, Mila (2024-2025) - VP Communications, CS Graduate Society (2022-2023) -> Managed society communications, introduced Instagram channel - VP Finance, CS Graduate Society (2021-2022) -> Architected website redesign, managed finances & events Entrepreneurship: - Co-president, Dalhousie Entrepreneurship Society - Stanford d.school University Innovation Fellow - Co-founder, Hoist Halifax (youth entrepreneurship program) - SHAD Fellow

============= EDUCATION ============= - Ph.D. in Computer Science, Université de Montréal 4.2/4.3 GPA) - M.Sc. in Computer Science, McGill (4.0 GPA) - B.Comp.Sci, Dalhousie (4.13 GPA, First Class Honours)

======================= CONTACT INFORMATION ======================= - Email: aristides.milios [at] mila.quebec - LinkedIn: linkedin.com/in/aristidesmilios - GitHub: github.com/amilios - Google Scholar: Google Scholar Profile - Twitter: @amilios

============== RESUME PDF ============== --2024-08-14 15:42:33-- https://atmilios.com/resume.pdf Resolving atmilios.com... 192.168.1.1 Connecting to atmilios.com:443... connected. HTTP request sent, awaiting response... 200 OK Length: 245760 (240K) [application/pdf] Saving to: 'resume.pdf' resume.pdf 100%[=============>] 240K 1.2MB/s in 0.2s 2024-08-14 15:42:34 (1.2 MB/s) - 'resume.pdf' saved [245760/245760] Click here to open resume.pdf

An intrepid ML researcher who thinks LLMs should know when to say "I don't know" 🤖 Currently teaching machines to be humble (with mixed results).

USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND aristides 1337 95.2 87.3 999999 42042 pts/0 R+ 09:00 47:23 neural_network_training --epochs=∞ aristides 1339 2.3 5.7 16384 2048 pts/2 S 14:15 1:42 paper_writing --deadline=yesterday aristides 1340 0.0 0.1 512 256 pts/3 S 15:30 0:01 debugging_life aristides 1341 15.7 12.4 32768 4096 pts/4 R 16:00 8:15 procrastinating --mode=reddit aristides 1342 1.1 3.2 8192 1536 pts/5 S 16:45 0:22 impostor_syndrome --silent root 1343 99.9 99.9 ∞∞∞∞∞ ∞∞∞∞∞ ? R 00:00 999:59 uncertainty_quantification