I'm Alberto Sinigaglia. Heres some infos about me:

• I'm doing a PhD in Deep Reinforcement Learning @ University of Padua
• I'm extremely proud of having published at ICML and Neurips
• very passioned on everything Deep Learning related
• very passioned on everything Reinforcement Learning related
• I have a Master Degree DataScience @ UniPD
• I have a Bachelor Degree in CompSci @ UniPD
• I really like helping people on StackOverflow and StackExchange

My very not updated personal website albertosinigaglia.me and my maybe a bit more updated CV

• Sinigaglia, A.*, Sartor, D.*, et al.
  Simple and Effective Specialized Representations for Fair Classifiers.
  Advances in Neural Information Processing Systems (NeurIPS 2025)

• Sinigaglia, A.*, Sartor, D.*, et al.
  Advancing Constrained Monotonic Neural Networks: Achieving Universal Approximation Beyond Bounded Activations.
  Forty-Second International Conference on Machine Learning (ICML 2025)

• Carletti, M., Sinigaglia, A., et al.
  On the Limitations of Adversarial Training for Robust Image Classification with Convolutional Neural Networks.
  Information Sciences (2024): 120703

• Sinigaglia, A., et al. Edge Delayed Deep Deterministic Policy Gradient: Efficient Continuous Control for Edge Scenarios.
  IEEE Transactions on Automation Science and Engineering (2025)

• Wiebe, F., Sinigaglia, A., et al.
  Reinforcement Learning for Robust Athletic Intelligence: Lessons from the 2nd "AI Olympics with RealAIGym" Competition.
  IEEE Robotics and Automation Magazine (2025)

* Equal contribution

I really appreciate spending time coding projects, out of which the following are the ones that I'm proud/I like the most:

• Force based graph: self organizing graph using non convex gradieng descet based optimization, demo online

• Binarized Neural Network: implementation of binarized neural networks using first order logic and MaxSAT solvers

• Snake using RL: Reinforcement learning agent that learns to play snake

• Block coordinate gradient descent methods: implementation of multiple block coodinate gradient descent algorithms for a convex high-dimensional loss function

• Variance reducing gradient descent algorihtms: implementation of SARAH, SNVRG, SpiderBoost, variance reducing algorithms for stochastic gradient descent optimization

• Hangman using RL: Reinforcement learning agent that learns to play hangman

• Attention sampling classification: classification of mega-pixel images using attention sampling, trained using Reinforcment learning

• Discrete GANs: GANs with a discrete generator, trained using Reinforcment Learning or Straight through estimator

• Pix2Pix: Pix2Pix implementation with multiple well-known network architectures

• Canvas animations (old): collections of animations done in pure JS using Canvas

  • webcam to checkboxes
  • webcam to text
  • lissajous
  • raycasting