Related papers: CSP-Atlas: Concept-Specific Neural Circuits in a S…
Transformers lack an explicit architectural mechanism for storing and organizing knowledge acquired during training. We introduce learnable sparse memory banks: a set of latent tokens, randomly initialized and trained end-to-end, that…
Encoding static images into spike trains is a fundamental step for enabling Spiking Neural Networks (SNNs) to process visual information. However, widely used methods such as rate coding, Poisson encoding, and time-to-first-spike (TTFS)…
We propose an Auto-Parsing Network (APN) to discover and exploit the input data's hidden tree structures for improving the effectiveness of the Transformer-based vision-language systems. Specifically, we impose a Probabilistic Graphical…
Neural networks often contain polysemantic neurons that respond to multiple, sometimes unrelated, features, complicating mechanistic interpretability. We introduce the Polysemanticity Index (PSI), a null-calibrated metric that quantifies…
Deep 'Analog Artificial Neural Networks' (ANNs) perform complex classification problems with remarkably high accuracy. However, they rely on humongous amount of power to perform the calculations, veiling the accuracy benefits. The…
Software-implementation, via neural networks, of brain-inspired computing approaches underlie many important modern-day computational tasks, from image processing to speech recognition, artificial intelligence and deep learning…
Transformer language models are neural networks used for a wide variety of tasks concerning natural language, including some that also require logical reasoning. However, a transformer model may easily learn spurious patterns in the data,…
The human brain represents the only known example of general intelligence that naturally aligns with human values. On a mere 20-watt power budget, the brain achieves robust learning and adaptive decision-making in ways that continue to…
Semantic Segmentation using deep convolutional neural network pose more complex challenge for any GPU intensive task. As it has to compute million of parameters, it results to huge memory consumption. Moreover, extracting finer features and…
High-performance computing systems are moving towards 2.5D and 3D memory hierarchies, based on High Bandwidth Memory (HBM) and Hybrid Memory Cube (HMC) to mitigate the main memory bottlenecks. This trend is also creating new opportunities…
The Vision Transformer (ViT) has achieved notable success in computer vision, with its variants widely validated across various downstream tasks, including semantic segmentation. However, as general-purpose visual encoders, ViT backbones…
Understanding how neural networks arrive at their predictions is essential for debugging, auditing, and deployment. Mechanistic interpretability pursues this goal by identifying circuits - minimal subnetworks responsible for specific…
Recently, significant improvement has been made on semantic object segmentation due to the development of deep convolutional neural networks (DCNNs). Training such a DCNN usually relies on a large number of images with pixel-level…
Reasoning segmentation requires models to ground complex, implicit textual queries into precise pixel-level masks. Existing approaches rely on a single segmentation token $\texttt{<SEG>}$, whose hidden state implicitly encodes both semantic…
Sparse autoencoders (SAEs) have become central to mechanistic interpretability, decomposing transformer activations into monosemantic features. Yet existing analyses characterise features almost exclusively through top-activating token…
Determining how the brain stores information is one of the most pressing problems in neuroscience. In many instances, the collection of stimuli for a given neuron can be modeled by a convex set in $\mathbb{R}^d$. Combinatorial objects known…
Transformers have demonstrated great success in numerous domains including natural language processing and bioinformatics. This success stems from the use of the attention mechanism by these models in order to represent and propagate…
Understanding the internal circuits that language models use to solve tasks remains a central challenge in mechanistic interpretability. A crucial part of finding circuits is understanding why each attention head attends where it does. To…
In this paper, we propose an easily trained yet powerful representation learning approach with performance highly competitive to deep neural networks in a digital pathology image segmentation task. The method, called sparse coding driven…
Qualitative reasoning involves expressing and deriving knowledge based on qualitative terms such as natural language expressions, rather than strict mathematical quantities. Well over 40 qualitative calculi have been proposed so far, mostly…