Gary B. Huang

Gary B. Huang, Viren Jain

Feedforward multilayer networks trained by supervised learning have recently demonstrated state of the art performance on image labeling problems such as boundary prediction and scene parsing. As even very low error rates can limit practical usage of such systems, methods that perform closer to human accuracy remain desirable. In this work, we propose a new type of network with the following properties that address what we hypothesize to be limiting aspects of existing methods: (1) a `wide' structure with thousands of features, (2) a large field of view, (3) recursive iterations that exploit statistical dependencies in label space, and (4) a parallelizable architecture that can be trained in a fraction of the time compared to benchmark multilayer convolutional networks. For the specific image labeling problem of boundary prediction, we also introduce a novel example weighting algorithm that improves segmentation accuracy. Experiments in the challenging domain of connectomic reconstruction of neural circuity from 3d electron microscopy data show that these "Deep And Wide Multiscale Recursive" (DAWMR) networks lead to new levels of image labeling performance. The highest performing architecture has twelve layers, interwoven supervised and unsupervised stages, and uses an input field of view of 157,464 voxels ($54^3$) to make a prediction at each image location. We present an associated open source software package that enables the simple and flexible creation of DAWMR networks.

Tongyuan Miao, Gary Huang, Kai Jun Han et al.

Diffusion Large Language Models (DLLMs) enable fully parallel token decoding but often remain impractical at inference time due to the many denoising iterations required to refine an information-free, fully masked initialization into coherent text. Most existing acceleration methods focus on traversing this generative trajectory more efficiently via improved solvers or sampling strategies. We advance a complementary perspective: shorten the trajectory itself by starting closer to the target distribution through context-aware initialization. We propose a training-free interface that injects prompt-conditioned priors from a lightweight auxiliary model into the diffusion initialization, and instantiate it with two mechanisms: discrete token injection and representation-level embedding interpolation. Because injected priors can be imperfect and unmask-only decoding can over-commit early, we also introduce a simple confidence-based remasking mechanism as a form of prior skepticism. Preliminary evidence on GSM8K suggests that context-aware initialization can substantially reduce denoising iterations (about 35\% fewer function evaluations in our setting), while also exposing a key open challenge: naive warm-starting can degrade final accuracy relative to strong diffusion baselines. We use these findings to motivate a research agenda around calibration, revision mechanisms, and representation alignment for reliable warm-started diffusion decoding.

Gary B Huang, Huei-Fang Yang, Shin-ya Takemura et al.

We propose a method to facilitate exploration and analysis of new large data sets. In particular, we give an unsupervised deep learning approach to learning a latent representation that captures semantic similarity in the data set. The core idea is to use data augmentations that preserve semantic meaning to generate synthetic examples of elements whose feature representations should be close to one another. We demonstrate the utility of our method applied to nano-scale electron microscopy data, where even relatively small portions of animal brains can require terabytes of image data. Although supervised methods can be used to predict and identify known patterns of interest, the scale of the data makes it difficult to mine and analyze patterns that are not known a priori. We show the ability of our learned representation to enable query by example, so that if a scientist notices an interesting pattern in the data, they can be presented with other locations with matching patterns. We also demonstrate that clustering of data in the learned space correlates with biologically-meaningful distinctions. Finally, we introduce a visualization tool and software ecosystem to facilitate user-friendly interactive analysis and uncover interesting biological patterns. In short, our work opens possible new avenues in understanding of and discovery in large data sets, arising in domains such as EM analysis.

Gary B. Huang, Louis K. Scheffer, Stephen M. Plaza

Extracting a connectome from an electron microscopy (EM) data set requires identification of neurons and determination of synapses between neurons. As manual extraction of this information is very time-consuming, there has been extensive research effort to automatically segment the neurons to help guide and eventually replace manual tracing. Until recently, there has been comparatively less research on automatically detecting the actual synapses between neurons. This discrepancy can, in part, be attributed to several factors: obtaining neuronal shapes is a prerequisite first step in extracting a connectome, manual tracing is much more time-consuming than annotating synapses, and neuronal contact area can be used as a proxy for synapses in determining connections. However, recent research has demonstrated that contact area alone is not a sufficient predictor of synaptic connection. Moreover, as segmentation has improved, we have observed that synapse annotation is consuming a more significant fraction of overall reconstruction time. This ratio will only get worse as segmentation improves, gating overall possible speed-up. Therefore, we address this problem by developing algorithms that automatically detect pre-synaptic neurons and their post-synaptic partners. In particular, pre-synaptic structures are detected using a Deep and Wide Multiscale Recursive Network, and post-synaptic partners are detected using a MLP with features conditioned on the local segmentation. This work is novel because it requires minimal amount of training, leverages advances in image segmentation directly, and provides a complete solution for polyadic synapse detection. We further introduce novel metrics to evaluate our algorithm on connectomes of meaningful size. These metrics demonstrate that complete automatic prediction can be used to effectively characterize most connectivity correctly.

Stephen M. Plaza, Toufiq Parag, Gary B. Huang et al.

Reconstructing neuronal circuits at the level of synapses is a central problem in neuroscience and becoming a focus of the emerging field of connectomics. To date, electron microscopy (EM) is the most proven technique for identifying and quantifying synaptic connections. As advances in EM make acquiring larger datasets possible, subsequent manual synapse identification ({\em i.e.}, proofreading) for deciphering a connectome becomes a major time bottleneck. Here we introduce a large-scale, high-throughput, and semi-automated methodology to efficiently identify synapses. We successfully applied our methodology to the Drosophila medulla optic lobe, annotating many more synapses than previous connectome efforts. Our approaches are extensible and will make the often complicated process of synapse identification accessible to a wider-community of potential proofreaders.

Gary B. Huang, Stephen Plaza

In this work, we propose a learning framework for identifying synapses using a deep and wide multi-scale recursive (DAWMR) network, previously considered in image segmentation applications. We apply this approach on electron microscopy data from invertebrate fly brain tissue. By learning features directly from the data, we are able to achieve considerable improvements over existing techniques that rely on a small set of hand-designed features. We show that this system can reduce the amount of manual annotation required, in both acquisition of training data as well as verification of inferred detections.

John A. Bogovic, Gary B. Huang, Viren Jain

For image recognition and labeling tasks, recent results suggest that machine learning methods that rely on manually specified feature representations may be outperformed by methods that automatically derive feature representations based on the data. Yet for problems that involve analysis of 3d objects, such as mesh segmentation, shape retrieval, or neuron fragment agglomeration, there remains a strong reliance on hand-designed feature descriptors. In this paper, we evaluate a large set of hand-designed 3d feature descriptors alongside features learned from the raw data using both end-to-end and unsupervised learning techniques, in the context of agglomeration of 3d neuron fragments. By combining unsupervised learning techniques with a novel dynamic pooling scheme, we show how pure learning-based methods are for the first time competitive with hand-designed 3d shape descriptors. We investigate data augmentation strategies for dramatically increasing the size of the training set, and show how combining both learned and hand-designed features leads to the highest accuracy.