Searcharxiv⌕ Search

arXiv subjects

Naeem Khoshnevis

Publications and source records attributed to Naeem Khoshnevis.

6 recordsLinked to original sources

Trainable sparse dendritic networks reveal regime-dependent roles for shunting inhibition and morphology

Neural circuits must compute with noisy inputs and adapt their responses to context. How dendritic structure and synaptic integration support these functions remains unclear. We introduce DendriNet, a trainable rate-based framework with sparse excitatory and inhibitory contacts, controllable dendritic architecture and alternative integration rules. In image-classification tasks, shunting networks transmit and rely on distal dendritic signals more strongly than additive networks. Distributing limited contacts across compartments improves classification in small populations, whereas greater serial dendritic depth provides no benefit on clean images within resource-matched trees. Robustness depends on input statistics: deeper shunting trees better tolerate independent pixel noise, while addition better tolerates strong global scaling. In a contextual task with gain variation and noise, shunting improves accuracy under the training corruption but depends less on context-specific inhibition than a separately tuned additive network. Learning leak conductances partly restores shunting accuracy on dim, noise-free inputs outside the training distribution. At AlexNet width, shunting integration and learned sparse connectivity support smaller deployed networks with accuracy comparable to retrained dense layers. These results distinguish the local action of synaptic integration from the computational roles acquired by inhibitory pathways through learning, and suggest experiments on dendritic function and design principles for brain-inspired networks.

q-bio.NC↗

GB-LSR: Local Spectral Decoding with a Learned Global Bandwidth for Arbitrary-Scale Super-Resolution

We present GB-LSR (Global-Bandwidth Local Spectral Representation), a fixed-grid local spectral representation for continuous image decoding. The image domain is partitioned into non-overlapping square patches. Each patch carries coefficients for a truncated Fourier basis, predicted by a single linear projection from shared convolutional-encoder features, and one trainable scalar bandwidth is shared across every patch and every image. As in earlier local spectral decoders, decoding at a continuous coordinate is a fixed-size basis contraction whose cost is set by the spectral cutoff; GB-LSR learns the bandwidth of that basis instead of fixing it. We evaluate an arbitrary-scale super-resolution extension, GB-LSR-Scalar-ASR, against the authors' released LIIF, LTE, and SRNO checkpoints on the same RDN encoder, with every method scored under one protocol and timed in one session per scale, each on one GPU. It runs 1.25x faster than LIIF-RDN at x4 and as fast as SRNO-RDN, whose released code uses 15 times as much peak memory on Urban100. It trails the three encoder-matched baselines by 0.07 to 0.79 dB PSNR-Y in distribution, SRNO-RDN by 0.35 dB on average. Removing the local ensemble raises the speedup to 2.41x over LIIF-RDN and 1.95x over SRNO-RDN at x4, and to 3.00x and 2.41x at x8, without changing PSNR-Y beyond seed variation, at the cost of value jumps at cell boundaries of 0.22 gray levels (of 255) on average at x4. Against the EDSR-baseline checkpoints of five recent methods at x4, GB-LSR-Scalar-ASR scores above or within 0.17 dB on PSNR-Y of LMF, SRNO-EDSR, and OPE-SR-EDSR (1.39 to 6.43 million parameters against 22.02) and 0.14 to 0.57 dB below GSASR and Thera (20.44 and 5.85 million), and has a higher mean LPIPS at x4 than every baseline.

cs.CV↗

When Branch-Local Shunting Helps: A Gain-Load-Alignment Principle for Dendritic E/I Networks

Biological neurons combine excitatory and inhibitory (E/I) activity on branched dendrites through shunting, in which inhibition divisively attenuates excitation. Whether this improves population readout over additive E/I integration of the same nonnegative inputs remains unclear. We introduce DendriNet, a trainable framework that varies integration rule, morphology, synaptic allocation, divisor locality, and dendritic nonlinearities. For population codes with multiplicative gain, a local linearization of any realizable shunting readout yields a decision direction within the positive additive E/I cone; matching the additive optimum requires a positive self-consistent shunting realization. Every scalar shunting threshold also has an exact affine additive realization. Beyond this local limit, performance follows a gain-load-alignment principle: branch-local shunting helps when a reliable divisor suppresses signal-aligned gain more than it attenuates signal or adds denominator variability. Passive additive trees flatten to linear readouts, whereas shunting trees compose local divisors. In a designed hierarchy, deep shunting outperforms tangent and fitted-linear controls, but flexible nonlinear predictors overtake it with enough labels. Support shuffling reverses the linear comparisons, sensor corruption reverses the fitted-linear comparison, and resource-matched activated training shows no consistent depth benefit. The same support and reliability interaction appears in frozen-feature normalization. Across three mouse V1 sessions, the shunting-over-additive decoder gap is largest for narrow readouts, reverses under strong private noise at the widest readout, and varies across running states. Morphology can determine where reliable nuisance estimates meet task-relevant signals, but neither depth nor shunting is intrinsically advantageous.

q-bio.NC↗

The Emergence of Complex Behavior in Large-Scale Ecological Environments

We explore how physical scale and population size shape the emergence of complex behaviors in open-ended ecological environments. In our setting, agents are unsupervised and have no explicit rewards or learning objectives but instead evolve over time according to reproduction, mutation, and selection. As they act, agents also shape their environment and the population around them in an ongoing dynamic ecology. Our goal is not to optimize a single high-performance policy, but instead to examine how behaviors emerge and evolve across large populations due to natural competition and environmental pressures. We use modern hardware along with a new multi-agent simulator to scale the environment and population to sizes much larger than previously attempted, reaching populations of over 60,000 agents, each with their own evolved neural network policy. We identify various emergent behaviors such as long-range resource extraction, vision-based foraging, and predation that arise under competitive and survival pressures. We examine how sensing modalities and environmental scale affect the emergence of these behaviors and find that some of them appear only in sufficiently large environments and populations, and that larger scales increase the stability and consistency of these emergent behaviors. While there is a rich history of research in evolutionary settings, our scaling results on modern hardware provide promising new directions to explore ecology as an instrument of machine learning in an era of increasingly abundant computational resources and efficient machine frameworks. Experimental code is available at https://github.com/jbejjani2022/ecological-emergent-behavior.

cs.MA↗

SpaCE: The Spatial Confounding Environment

Spatial confounding poses a significant challenge in scientific studies involving spatial data, where unobserved spatial variables can influence both treatment and outcome, possibly leading to spurious associations. To address this problem, we introduce SpaCE: The Spatial Confounding Environment, the first toolkit to provide realistic benchmark datasets and tools for systematically evaluating causal inference methods designed to alleviate spatial confounding. Each dataset includes training data, true counterfactuals, a spatial graph with coordinates, and smoothness and confounding scores characterizing the effect of a missing spatial confounder. It also includes realistic semi-synthetic outcomes and counterfactuals, generated using state-of-the-art machine learning ensembles, following best practices for causal inference benchmarks. The datasets cover real treatment and covariates from diverse domains, including climate, health and social sciences. SpaCE facilitates an automated end-to-end pipeline, simplifying data loading, experimental setup, and evaluating machine learning and causal inference models. The SpaCE project provides several dozens of datasets of diverse sizes and spatial complexity. It is publicly available as a Python package, encouraging community feedback and contributions.

cs.LG↗

CausalGPS: An R Package for Causal Inference With Continuous Exposures

Quantifying the causal effects of continuous exposures on outcomes of interest is critical for social, economic, health, and medical research. However, most existing software packages focus on binary exposures. We develop the CausalGPS R package that implements a collection of algorithms to provide algorithmic solutions for causal inference with continuous exposures. CausalGPS implements a causal inference workflow, with algorithms based on generalized propensity scores (GPS) as the core, extending propensity scores (the probability of a unit being exposed given pre-exposure covariates) from binary to continuous exposures. As the first step, the package implements efficient and flexible estimations of the GPS, allowing multiple user-specified modeling options. As the second step, the package provides two ways to adjust for confounding: weighting and matching, generating weighted and matched data sets, respectively. Lastly, the package provides built-in functions to fit flexible parametric, semi-parametric, or non-parametric regression models on the weighted or matched data to estimate the exposure-response function relating the outcome with the exposures. The computationally intensive tasks are implemented in C++, and efficient shared-memory parallelization is achieved by OpenMP API. This paper outlines the main components of the CausalGPS R package and demonstrates its application to assess the effect of long-term exposure to PM2.5 on educational attainment using zip code-level data from the contiguous United States from 2000-2016.

stat.CO↗