Multi-Layer Perceptron & Backpropagation — Interactive Explorer Interactive explainer for the multi-layer perceptron: XOR, hidden-layer representations, backpropagation as gradient flow, and depth vs. width as a complexity story.

After getting some questions from students about MLP architectural choices and how backpropagation works, I put together this comprehensive multi-layer perceptron/backpropagation explorer (with a lot of cool latent-space visualizations, some in 3D!):
tpavlic.github.io/asu-bioinspi...

Spiking Neural Networks — interactive explorer Interactive explorer for spiking neural networks: simulate leaky integrate-and-fire neurons, visualize spike trains and membrane potentials, and explore Spike-Timing-Dependent Plasticity (STDP) — the ...

Ok. I think the last widget I'll be building for my Bio-Inspired #AI and Optimization course this semester.

Spiking Neural Network ( #SNN ) explorer

LIF Neuron model, Rate coding, STDP, Survey of HW&SW implementations, #Neuromorphic applications & learning
tpavlic.github.io/asu-bioinspi...

Toward Multimodal AI — interactive explainer Interactive explainer tracing the path from CNNs and patch embeddings to Vision Transformers (ViT), CLIP-style contrastive pretraining, and modern multimodal architectures — showing how a single atten...

Put together this "Toward Multimodal #AI" explainer widget. Starts with a motivation for why #ViT's over CNN's, introduces joint embedding spaces (like #CLIP), and then shows how to adapt those spaces as inputs to #LLM's for true multimodal reasoning.
tpavlic.github.io/asu-bioinspi...

On Tab 3, about self-supervised training of transformers, it makes connections with Latent Learning in psychology and the debate about whether latent learning is actually associative learning (similar to whether some self-supervision is actually just supervision, as in GPT).

Transformer Architecture Explorer — Bio-Inspired AI & Optimization Interactive explorer for the Transformer architecture: step through scaled dot-product self-attention, multi-head attention, positional encodings, and Vision Transformers (ViT) with live visualization...

Students have a lot of questions and misconceptions about transformers. Here's an interactive explainer I built that covers transformers (3 types – Enc only, Dec only, Enc–Dec), attention/cross-attention, self-supervision, and Vision Transformers (ViT) too.
tpavlic.github.io/asu-bioinspi...

Autoencoder Explorer — interactive explainer Train a deep autoencoder on MNIST-like digit data and watch the 2-D bottleneck encoding cluster by class — a hands-on demonstration of unsupervised representation learning with neural networks.

Another interactive AI educational tool. The basic autoencoder is a classic example of unsupervised learning that (in contrast with Hebbian updating) leverages mechanisms of supervised training in neural networks. In-browser clustering in MNIST-like task:
tpavlic.github.io/asu-bioinspi...

New Recurrent Neural Network (RNN) Explorer I built for my students.

Tabs: TDNN's (think FIR), Output feedback (IIR), RNN's with autoregressive feedback of latent states (with randomized reservoir support), BPTT training explorer, then LSTM/GRU explainer.
tpavlic.github.io/asu-bioinspi...

Everyone is forgetting these marine workers trapped around the strait.

That unsupervised learning memristor crossbar example from Boyn et al. (2017) can also be implemented with ANN's with Hebbian updating and latent inhibition. I've put together a teaching demo widget [inspired by Rumelhart and Zipser (1985)] of this at:
tpavlic.github.io/asu-bioinspi...

#ProTip: The relatively new "LaTeX to Math" in #PowerPoint lets you highlight #LaTeX within math mode and convert it into PPT's #math format.

Someone with the handle @velocity69.bsky.social is sharing a video with the caption:

"The highest quality video of the moon was just released…
so beautiful

* Artemis II"

It's gone viral, but it's not from Artemis II. If you've shared it, you might want to un-share.

Ferroelectric Memristor Synapses & Crossbar Learning — interactive explainer Explore how ferroelectric memristor synapses in a crossbar array implement Spike-Timing-Dependent Plasticity (STDP) — adjust pulse timings, device parameters, and network architecture to watch Hebbian...

Boyn et al. (2017) described a crossbar memristor array that could do unsupervised learning of patterns (e.g., from MNIST image data) purely in hardware using spike-timing-dependent plasticity (STDP) of the memristors. Here's a teaching demo of that result
tpavlic.github.io/asu-bioinspi...

Happy #FirstContactDay for those who celebrate! 🖖

Reservoir Computing — Echo State Network Explorer Interactive reservoir computing explorer using Echo State Networks: adjust spectral radius, sparsity, and input scaling to watch a fixed random reservoir map inputs to separable dynamics.

Reservoir Computing is traditionally a difficult topic for students -- using an untrained recurrent neural network as a feature extractor for time-series analysis. I put together this widget to highlight how these tools turn temporal into spatial patterns.
tpavlic.github.io/asu-bioinspi...

Single-Layer Perceptron — Neuron & Lever Explainer Animated explainer linking the biological neuron to the linear classifier: adjust synaptic weights and inputs to watch the decision boundary shift in real time.

In my Bio-Inspired AI survey course, I'll be starting the unit on neural computing – a large unit going from SLP's up through LLM's. Here's a little Single-Layer Perceptron demo I built demonstrating why we call neural network parameters "weights".
tpavlic.github.io/asu-bioinspi...

MVT Explorer — Marginal Value Theorem — interactive explainer Interactive explorer for Charnov's Marginal Value Theorem (MVT): drag the tangent to find optimal patch residence time and see how the habitat-wide rate R* acts as a temporal discount rate — a structu...

For the behavioral ecologists, here's an interactive Marginal Value Theorem (MVT) explorer for classical single-patch case [Tab 1] as well as the multi-patch "temporal portfolio" case [Tab 2]. Good for teaching Optimal Foraging Theory or opportunity cost.
tpavlic.github.io/asu-bioinspi...

Lyapunov Level-Set Explorer — interactive explainer Interactively explore Lyapunov functions and level-set curves for common nonlinear dynamical systems to build intuition about Lyapunov stability.

Just to demo how I might use frontend tools built with #ClaudeAI in a controls class, I put together this Lyapunov function candidate/phase portrait explorer. I like it! I think this would be a helpful tool in an nonlinear systems class.
tpavlic.github.io/topic_visual...

There are people typing this into ChatGPT right now to get an answer.

That's a very big -log(p) value! Can't argue with that.

Monte Carlo Explorer — interactive explainer Interactive explorer for Monte Carlo methods: visualize rejection sampling, Monte Carlo integration, and Markov Chain Monte Carlo to build intuition for randomized algorithms.

I was explaining Monte Carlo integration today in class, and I realized I could use a new widget. So, here it is. Tab 1 is simple MC estimation of pi. Tab 2 is MC integration ("random Riemann rectangles") for multiple functions. Conf. interval review too!
tpavlic.github.io/asu-bioinspi...

Ant Foraging Dynamics Explorer — interactive explainer Explore how ants collectively select foraging trails through pheromone-based stigmergy: experiment with trail noise and Y-maze decision-making, then see how recruitment linearity shapes collective pat...

And if you like #ants, I also put together this Ant Foraging Dynamics Explorer that explores how noisy and/or highly localized, one-on-one communication can make ants (and bees) more responsive to dynamic environments. Check out tabs 1 and 2 for details.
tpavlic.github.io/asu-bioinspi...

ACO / Ant System Explorer — interactive explainer Explore Ant Colony Optimization: watch a pheromone-guided ant colony solve a layered combinatorial problem and a Travelling Salesman Problem, and see how stigmergic reinforcement drives collective sea...

Another Swarm Intelligence demo put together in the nick of time for my Bio-Inspired AI and Optimization course this semester. This time Ant Colony Optimization (ACO), with both a layered DAG decision-variable encoding and the traditional TSP. Try it out!
tpavlic.github.io/asu-bioinspi...

A gentle introduction to swarm optimization and related topics, in section 6.1 of our book. ehudlamm.com/biocomputing...

Particle Swarm Optimization (PSO) Explorer — interactive explainer Explore Particle Swarm Optimization: step through the velocity-update rule with cognitive and social components, then run a live swarm on a 2-D landscape and watch particles converge on optima.

The excitement about collective motion at that time gave rise to an optimization metaheuristic that is still widely used today – Particle Swarm Optimization (PSO). Here's an interactive PSO explorer that shows how these models can search complex decision spaces.
tpavlic.github.io/asu-bioinspi...

Vicsek Model Explorer — interactive explainer Explore the Vicsek model of collective motion: adjust noise and interaction radius to watch self-propelled particles transition between disordered flocking and coherent swarm behavior.

And for comparison, here's a similar interactive tool of the Vicsek collective-motion model that has fewer macroscopic phases but a little simpler model. A phase diagram and presets are also included for easy exploration of the different modes.
tpavlic.github.io/asu-bioinspi...

Boids Explorer — interactive explainer Explore Reynolds' Boids model of collective motion: tune separation, alignment, and cohesion weights to watch emergent flocking arise from three simple local rules — a classic inspiration for swarm in...

And now some collective-motion interactive demonstration tools for the classic "swarming" motion models. First up, Reynolds' "Boids" with a visualizer including a phase diagram based on @icouzin.bsky.social's 2002 JTB paper.
tpavlic.github.io/asu-bioinspi...

Electronic passport renewal is amazing. Routine service maybe took 15 days! Photo from a cell phone. Highly recommended. Remember that you may soon need passport to vote (particularly if name doesn't match birth certificate), and a Real ID isn't proof of citizenship.

Artificial Immune Systems Explorer — interactive explainer Explore two core artificial immune system algorithms side by side — Negative Selection (NSA) trains detectors on self-data to flag anomalies, while CLONALG uses clonal selection and affinity maturatio...

One more bio-inspired widget built using #ClaudeAI's frontend skills. This one showcases two classic Artificial Selection Algorithms – Negative Selection Algorithm (NSA) and CLONALG – immunologically inspired pattern matchers for anomaly/threat detection.
tpavlic.github.io/asu-bioinspi...

Genetic Algorithm Explorer — interactive explainer Explore genetic algorithm dynamics across four interactive tabs: a 1-D multimodal landscape with niching, island model / distributed GAs, a 2-D optimization zoo, and a two-objective MOEA with Pareto r...

It was a lot of fun guiding #ClaudeAI to build this Genetic Algorithm (GA) Explorer, which allows for investigating single-objective GA's in numerous configurations (including multi-modal and island GA's) and multi-objective GA's. Check out all 4 tabs!
tpavlic.github.io/asu-bioinspi...

Divisive ‘Star Trek: Starfleet Academy’ Canceled After Two Seasons The Paramount+ 'Star Trek' show will end after its already filmed second season.

Starfleet Academy sadly was not a recipe for Star Trek living long and prospering.

"Divisive ‘Star Trek: Starfleet Academy’ Canceled After Two Seasons"
www.hollywoodreporter.com/tv/tv-news/s...