Arnold Pears’ Post

"Can I play a simulation more than once?" Watch here as Ken Spero, Founder and President of SchoolSims, discusses the power of playing through a simulation multiple times. https://lnkd.in/enge6TiT #Edchat #EdTech #BackToSchool #Simulation #ProfessionalDevelopment

📢📢📢Introduce our #ICML workshop work: Fréchet Video Motion Distance: A Metric for Evaluating Motion Consistency in Videos TL:DR: We propose the Fréchet Video Motion Distance (FVMD), a novel metric to evaluate the motion consistency of video generation. 📜ArXiv: https://lnkd.in/gQ9fz-X4  💻Code: https://lnkd.in/gm-CzKwV  📝Blog: https://lnkd.in/gxetDDiN Our FVMD pipeline first tracks video key point trajectories using the pre-trained PIPs++ model and computes the velocity and acceleration fields for each frame. The motion features are then derived from the histograms of the quantized velocity and acceleration. FVMD is eventually given by the Fréchet distance between the motion features of generated and ground-truth videos. Joint work with Youran Qu, Qi YAN, Xiaohui Zeng, Lele Wang, Renjie Liao

New experiment with freeform General Circle Coordinates correlation based computational axes arrangement. #DataAnalysis #Visualization #DataVisualization #CoordinateSystems

Day-12 : Max Circular Subarray Sum Solving hard problems like the 'Max Circular Subarray Sum' makes you a better programmer! This one tested my understanding of array manipulation and optimization. Combining Kadane’s Algorithm for linear subarrays with tricks for circular ones was a game-changer. Keep learning, keep growing #CodingLife #AlgorithmDesign #GrowthMindset #GfG #160DayChallenge #WomenInTech

Latest experiments in guided generation This toy algorithm is an example of a functionality which could be adapted to greatly improve guided generation in LLMs. Currently employing a completion model, it can be seen as one method to rely on LLM output while ensuring a desirable generation.

Excited to share our NeurIPS paper UniMTS: a unified pre-trained model for motion time series / human activity recognition! 🚀 Existing motion time series classification / HAR models are trained and tested on the same dataset, leading to poor generalizability across variations in device location, device mounting orientation and human activity type. With UniMTS, we introduce the first unified pre-training procedure for motion time series that generalizes across diverse device latent factors and activities with impressive zero-shot, few-shot and full-shot performance. 📃Paper: https://lnkd.in/gWJCtj4V 💻Code: https://lnkd.in/gdsiBYpu 🤖Model: https://lnkd.in/g9wXCc3u #timeseries #har #motion #llm

A Remarkable Experience in Molecular Docking! I had the incredible opportunity to participate in a hands-on Molecular Docking Online Workshop conducted by Research Boulevard. This workshop was a comprehensive deep dive into the world of molecular docking, offering a detailed, step-by-step guide using tools like AutoDock4, OpenBabel, MGL Tools, and Maestro Viewer. During the session, I gained practical knowledge on: - Setting up and performing molecular docking using AutoDock4. - Conducting protein-ligand interaction analysis to understand binding mechanisms. - Utilizing advanced tools for visualization and streamlining workflows. It was a fantastic blend of theory and hands-on practice, making complex concepts approachable and actionable. #MolecularDocking #AutoDock #ProteinLigandInteractions #DrugDiscovery #ContinuousLearning #computationalbiology

"Exploring Motion with a Ticker Timer: In this experiment, we used a ticker timer to measure motion by analyzing the dots on a paper tape. First, we calibrated the timer by pulling the tape uniformly. Then, students explored different motion patterns by pulling the tape freely. Finally, we connected the tape to a trolley and let it coast down an inclined plane, capturing the motion in real-time. Watch as we break down the fundamentals of motion, step by step!" #MotionExperiment #TickerTimer #PhysicsInAction #LearnPhysics #ScienceExperiments #PhysicsMadeEasy #STEMLearning #InclinedPlane #RealTimeMotion #HandsOnLearning

I also start sharing the train and eval codes & weights for `Cross-D Conv: Cross-Dimensional Transferable Knowledge Base via Fourier Shifting Operation`