Farhad Maleki

Failure prediction based on execution logs using Deep Learning Congratulations to Fatemeh (Bahar) Hadadi (University of Ottawa, Canada) for getting the first (large) paper of her PhD work accepted in the Empirical Software Engineering Journal (Springer): “Systematic Evaluation of Deep Learning Models for Failure Prediction", Bahar Hadadi, Joshua Dawes, Donghwan Shin, Domenico Bianculli, and Lionel Briand. This is about systematically investigating and understanding the factors that affect the accuracy of various deep learning strategies for predicting failures from execution logs. Results are rather surprising ... As in many other cases, simpler is better. Preprint: https://lnkd.in/eDxJXusK This is part of a collaboration project between The University of Luxembourg (SnT) and the Nanda lab at uOttawa.

39

🗣️ Tomorrow, at 13:00, our PhD Student Vahid Ghafouri will give a talk in our #NetworksWeeklySeminar. 🔹During his seminar, he will discuss the impact of echo chambers on social media, proposing an unsupervised metric using Transformer models to quantify echo chamber effects, focusing on user diversity and polarization in controversial and non-controversial topics. 👉 All the details about the event: https://lnkd.in/eh52BeDt

29

We invite research submissions on Quantum Computing and Reinforcement Learning (QCRL) at IEEE Quantum Week 2024 (Montreal, Sep. 15-20, 2024). https://2024.qcrl.io/ Scope and Topics: Topics relevant to this session include, but are not limited to: - Theoretical studies of quantum reinforcement learning algorithms - Challenges in quantum reinforcement learning in the NISQ era - Quantum reinforcement learning under the influence of quantum device noise QRL in the context of trustworthy ML (e.g., adversarial attacks, federated learning) - Quantum reinforcement learning for scientific discovery, commercial, and industrial applications - Benchmarking QRL algorithms - Reinforcement learning for quantum control, quantum error correction, and quantum error mitigation - Reinforcement learning for quantum architecture search, optimization, and program synthesis - Hybrid RL-QC systems design Important Deadlines: - Jul 10, 2024: Abstract Due Date - Jul 15, 2024: Paper Submission Due Date - Jul 28, 2024: Notification of Paper Acceptance - Aug 5, 2024: Camera-ready Papers Due #quantum #machinelearning #reinforcementlearning

29

I will be teaching an ICPSR Summer Workshop, "Sensitivity Analysis: Quantifying the Robustness of Inferences to Alternative Factors or Data", Hybrid (virtual, FtF in Ann Arbor, MI, with asynchronous access to materials). https://lnkd.in/etB9AR8h Most social scientists need to consider the unobserved conditions that could change an inference. This workshop is about the statistical tools (R, Stata or the konfound-it app https://lnkd.in/ep_z-DJS ) to do just that, quantifying the strength of evidence to inform policy, practice, or general social science. The workshop is a full 20 instruction hours of instruction including opportunities to adapt to specific models such as mediation, moderation, diff in diff, logistic regression, hazard/cox regression, multilevel models, differential attrition from treatment groups, etc. Go here to register: https://lnkd.in/etB9AR8h Group discounts available as well. https://lnkd.in/eWRv8G7q

10

🎉 Congratulations to Mohammad Mahdi Johari from the Idiap Research Institute Laboratory on the successful defense of his PhD thesis “Advancing Self-Supervised Deep Learning for 3D Scene Understanding” under the supervision of Prof. François Fleuret and Prof. Daniel Gatica-Perez. 🔎 His research uses deep learning to tackle key challenges in 3D computer vision, such as depth estimation, novel view synthesis, and SLAM. The thesis proposes novel methods for each task that achieve state-of-the-art performance while being more efficient or requiring less data to train. #epfl_edee

20

There's only 5 days left to submit your paper! Remember, we only a short 1-page abstract

2

🌟 Reconnecting with a long-standing Friend and Discussing the Future of AI in Computer Vision & Medical Image Analysis 🔍I recently had the incredible opportunity to host Professor Tal Arbel and her students from McGill University and Mila - Quebec Artificial Intelligence Institute at TheraPanacea. Tal is not only an outstanding colleague and a great asset to the community but also a fantastic friend. Her energy, motivation, and engagement extend far beyond the scientific realm.  🩺 We discussed how AI and machine learning have revolutionized medical image analysis, enhancing diagnostic accuracy and predictive analytics. Technologies like deep learning are transforming healthcare, leading to better patient outcomes and more efficient workflows in fields such as oncology and neurology. 🚀 Our primary focus, however, was on future directions. We explored concepts like personalized medicine, which aims to tailor treatments based on individual imaging data, moving towards holistic treatment pathways. We also emphasized the importance of collaborative efforts in an academic-industry startup setting, involving clinical experts, advanced algorithms, and data management to drive innovation. The synergy between academia and industry will be pivotal in shaping the future of healthcare. #AI #MedicalImaging #HealthcareInnovation #FutureOfMedicine #AcademicConnections

108

5 Comments

I am very happy that our paper "An inexact column-and-constraint generation method to solve two-stage robust optimization problems" continues to attract attention as one of the most downloaded papers of ORL. Kudos to my exceptional PhD student, Man Yiu Tsang, who led this work, and to my fantastic colleague, Prof. Frank E. Curtis, for his invaluable contribution. Man Yiu Tsang will soon be in the academic job market. https://lnkd.in/eSsE-ZEm #RobustOptimization #DistributionallyRobustOptimization #Algorithms #Operationsresearch #Optimzation

92

1 Comment

In our latest paper with Jean-Baptiste Waring, B.Eng., we generated a dataset on `ibm_quebec` the 127-qubit IBM Quantum System One located in Canada. The dataset contains more than 4000 experiments of long-range entanglement path qubits from 0 to 100+ qubits. The dataset and the code can be found here: https://lnkd.in/eussBW9G Paper: https://lnkd.in/ekb9Vqja A special congratulations to Jean-Baptiste Waring, B.Eng. for this repo and the quality of the experiments. #quantumcomputing

47

1 Comment

I'm thrilled to announce that my latest research paper on melanoma detection has been published in SAGE Journals! (Farrukh A. Khan, PhD, FBCS, SMIEEE, Mamoona HUmayun) Discover how our approach enhances privacy and accuracy in medical imaging. Read it here: https://lnkd.in/eckYtskQ I look forward to your thoughts and feedback! #Research #MelanomaDetection #HyperParameterTuning #CNN #MedicalImaging #AcademicPublishing

21

7 Comments

Infini-Transformers: Efficient Infinite Context Modeling with Bounded Memory Munkhdalai et al. introduce Infini-Transformers, a novel approach that enables Transformer-based Large Language Models (LLMs) to process infinitely long inputs with bounded memory and computation. The key component is the Infini-attention mechanism, which incorporates a compressive memory into the standard attention mechanism and combines both masked local attention and long-term linear attention in a single Transformer block. Infini-attention reuses the query, key, and value states from the dot-product attention computation for long-term memory consolidation and retrieval. This allows for efficient plug-and-play long-context adaptation and speeds up training and inference. The compressive memory is parameterized with an associative matrix, and the memory update and retrieval process is cast as a linear attention mechanism. Experiments demonstrate the effectiveness of Infini-Transformers on long-context language modeling benchmarks, outperforming baseline models while achieving a 114x compression ratio. The approach also enables a 1B LLM to scale to 1M sequence length and solve the passkey retrieval task. Furthermore, an 8B model with Infini-attention achieves state-of-the-art results on a 500K length book summarization task after continual pre-training and fine-tuning. Infini-Transformers introduce minimal bounded memory parameters and enable fast streaming inference for LLMs. The subtle but critical modification to the Transformer attention layer allows LLMs to process infinitely long contexts with bounded memory and computation resources, paving the way for more efficient and scalable long-context modeling. Reference: https://lnkd.in/gxEcD7G4

4

📢 Join us on May 24th for a compelling webinar titled "Understanding and Navigating Bias in Machine Learning Models" with Dr. Afaf Taïk, a postdoctoral researcher at Mila and Université de Montréal, specializing in responsible AI with a focus on fairness and privacy. ✂ Despite the potential of machine learning models, they present numerous challenges, particularly regarding their biases, which can adversely affect already marginalized communities and individuals. For instance, text-to-image generative models have been shown to perpetuate and intensify gender and racial stereotypes. These concerning risks have led to the development of the field of fairness in ML, which seeks to identify, measure, and mitigate such biases. 📚 During this webinar, Dr. Afaf will delve into how biases emerge and spread throughout the ML pipeline. We will explore the approaches taken by the fairness in ML field to address these sociotechnical issues, discuss various mitigation techniques, and stress the necessity of implementing responsible practices to ensure fairness throughout the ML pipeline. 📅 Date: Wednesday, May 24th 🕗 Time: 8 PM (GMT+1), 9 PM (CET) Please take a minute to RSVP to receive the event Zoom link and mark your calendar. 📝 Register here: https://lnkd.in/e9qbFfpJ We look forward to having you join us for this insightful session! #MachineLearning #BiasInML #ResponsibleAI #FairnessInML #Privacy #AIethics #Webinar #TechTalk #AIResearch

88

I would like to share our recent work on various aspects of building explainable AI (XAI) systems. We tried to provide a comprehensive overview of XAI system development in collaboration with my former students at Shiraz University, Mohammad Mahdi Sayyadnejad and Ali A. (currently at the University of Naples Federico II) and former colleague Dr Hooman Tahatori by analysing developers' discussions titled "Exploring the Black Box: Analyzing Explainable AI Challenges": https://lnkd.in/dsC7xmYH 🔍 The study addresses the following research questions: RQ1: What topics are XAI developers asking about? RQ2: What Kind of questions are XAI developers asking? RQ3: What topics exhibit higher levels of popularity and difficulty? RQ4: Which technologies are XAI Developers Using? RQ5: How do the XAI Topics Evolve? I look forward to your comments and feedback. #XAI #ArtificialIntelligence #MachineLearning #DataScience #EmpiricalSoftwareEngineering #DevelopersDiscussions

69

Predictions based on Machine Learning and Statistics often rely on maximal likelihood, which can trap LLMs in biases inherited from the majority opinions of training data. This problem has been substantially addressed through conditional statistics, allowing for mitigation of biases and preventing the cancellation of minority perspectives. https://lnkd.in/gMXKqwyB

7

Alternators For Sequence Modeling Mohammad Reza Rezaei, Adji Bousso Dieng Abstract This paper introduces alternators, a novel family of non-Markovian dynamical models for sequences. An alternator features two neural networks: the observation trajectory network (OTN) and the feature trajectory network (FTN). The OTN and the FTN work in conjunction, alternating between outputting samples in the observation space and some feature space, respectively, over a cycle. The parameters of the OTN and the FTN are not time-dependent and are learned via a minimum cross-entropy criterion over the trajectories. Alternators are versatile. They can be used as dynamical latent-variable generative models or as sequence-to-sequence predictors. When alternators are used as generative models, the FTN produces interpretable low-dimensional latent variables that capture the dynamics governing the observations. When alternators are used as sequence-to-sequence predictors, the FTN learns to predict the observed features. In both cases, the OTN learns to produce sequences that match the data. Alternators can uncover the latent dynamics underlying complex sequential data, accurately forecast and impute missing data, and sample new trajectories. We showcase the capabilities of alternators in three applications. We first used alternators to model the Lorenz equations, often used to describe chaotic behavior. We then applied alternators to Neuroscience, to map brain activity to physical activity. Finally, we applied alternators to Climate Science, focusing on sea-surface temperature forecasting. In all our experiments, we found alternators are stable to train, fast to sample from, yield high-quality generated samples and latent variables, and outperform strong baselines such as neural ODEs and diffusion models in the domains we studied. 👉 https://lnkd.in/df6CadEk #machinelearning

1

Our article on 𝐒𝐜𝐫𝐞𝐞𝐧𝐢𝐧𝐠 𝐀𝐫𝐭𝐢𝐜𝐥𝐞𝐬 𝐟𝐨𝐫 𝐒𝐲𝐬𝐭𝐞𝐦𝐚𝐭𝐢𝐜 𝐑𝐞𝐯𝐢𝐞𝐰𝐬 𝐰𝐢𝐭𝐡 𝐂𝐡𝐚𝐭𝐆𝐏𝐓, co-authored with Eugene Syriani and Gauransh K., has been accepted for publication in the Journal of Computer Languages (COLA). This is the 𝐟𝐢𝐫𝐬𝐭 𝐝𝐞𝐭𝐚𝐢𝐥𝐞𝐝 𝐞𝐯𝐚𝐥𝐮𝐚𝐭𝐢𝐨𝐧 of #ChatGPT's ability to support one of the most labor-intensive and error-prone activities in #empirical #research. A comparison of ChatGPT's performance to that of five other classifiers, in six prompting strategies, six metrics, and five real large-scale data sets from systematic literature reviews and mapping studies. Takeaways: ➕ ChatGPT outperforms traditional classifiers (up to 82% accuracy), and that ➕ without prior training; however, ☝️ human intelligence is still required. ➡ Anticipate the next generation of #systematic #review tools to integrate #LLMs and human effort. Skim through our article for leads and details. Pre-print available: https://lnkd.in/d5xxt9Jn

21

Last Friday we hosted an enlightening webinar on "Understanding and Navigating Bias in Machine Learning Models" presented by Dr. Afaf Taïk, a postdoctoral researcher at Mila and Université de Montréal, specializing in responsible AI with a focus on fairness and privacy. During this webinar, Dr. Afaf Taïk delved into the categories of harms caused by unfair machine learning models, exploring the reasons behind these unfairness, which can stem from biased models or biased data. Additionally, Dr. Taïk examined various approaches to defining fairness in machine learning, emphasizing the importance of considering the consequences and context of the model's use case. Finally, Dr. Taïk discussed several techniques to avoid bias and achieve fairness, presenting the pros and cons of each strategy. 🎥 If you missed the webinar or would like to watch it again, please check the link to the recording in the comment section below. We extend our heartfelt thanks to Dr. Afaf Taïk for sharing her expertise and to all attendees for the insightful discussions at the end of the webinar. #MoroccoAI #Webinar #ArtificialIntelligence  #BiasInML #AIFairness

35

1 Comment

Our CVPR 2024 paper “Don’t Look into the Dark: Latent Codes for Pluralistic Image Inpainting” (https://lnkd.in/gY82ccew) is the perfect resource for your image inpainting task! Our model includes a carefully designed partial encoder that focuses solely on the visible part and a bidirectional transformer that infers missing labels. This allows our model to generate pluralistic images, even under extreme mask conditions! We've released the code and model here: https://lnkd.in/gQBefa7T. Explore the model and talk to us at CVPR! #cvpr2024

19