Gradient Descent on Two ReLU Neurons: Global Landscape and Bifurcation Dynamics
Authors: Binghua Li, Mengzhe Li, Denny Wu, Tianhao Wang
Organizations: University of California, Berkeley · École Normale Supérieure Paris-Saclay · New York University & Flatiron Institute · University of California, San Diego
Understanding how gradient descent learns features remains a central challenge in neural network theory. We study this question in a minimal multi-index setting that already exhibits rich multi-phase gradient dynamics: a well-specified two-neuron ReLU teacher-student model under isotropic Gaussian inputs, trained by population gradient flow from small initialization. We show that the dynamics are organized by two directions: the "easy" bisector direction, which carries the leading signal, and the "hard" splitting direction, which governs specialization. To characterize the loss structure and learning behavior, we analyze the population squared-loss landscape and show that every critical point with both student weights nonzero is either a global minimum or a saddle. We then track the gradient-flow trajectory, showing that student neurons first collapse toward a bisector saddle before escaping and specializing to teacher neurons. Our results provide both a landscape and dynamics account of the multi-phase symmetry-breaking behavior that arises in a simple multi-index model and under standard algorithmic and architectural choices.
Figures & tables
Figure 1: Illustration of bifurcation dynamics.
Figure 2: Escape time Tε and global-convergence time Tglobal vs. ambient dimension d , one panel per α , for random initialization at fixed σ=10−4 . Each point is the mean over 5 seeds, and error bars are ±1 s.d. The time Tε is when the total squared distance of the two student weights from the bisector axis, ∑i=12∑j=2dzi,j(t)2 , first reaches ε2 ( ε=0.1 ); Tglobal is when the student state first reaches distance 0.05 from the global minimizer set. Dashed lines are log- d fits; fitted slopes ∈[0.48,0.55] match the 21logd prediction above, and the gap Tglobal−Tε shown in the panel titles is almost d -independent.
Neural network training dynamics studies how parameters and representations change during optimization and how those changes shape learning, stability, and generalization.
In this paper, we study the gradient descent dynamics for jointly training both layers of a one-hidden-layer ReLU network to fit a linear target function. Concretely, we consider a realizable setting where inputs are drawn i.i.d. from a Gaussian distribution and labels follow a planted linear model. This stylized framework captures salient features of end-to-end training in inverse problems and certain auto-encoder models. Despite its apparent simplicity, the dynamics remain poorly understood, in part because the loss landscape contains multiple non-strict saddle points, making it unclear why gradient descent from random initialization reliably escapes bad stationary regions. We provide a detailed characterization of the optimization landscape and prove that gradient descent from a moderately small random initialization-simultaneously training both layers-converges to a global minimizer at a linear rate with order-wise optimal sample complexity. Our analysis tracks the trajectory through three phases: an alignment phase in which hidden weights progressively align with the planted direction while the output weights maintain the correct sign pattern; a growth phase in which the norms of both layers increase while preserving alignment; and a local refinement phase in which the aligned neurons rapidly converge to the planted direction, yielding fast local convergence. To rigorously show that GD avoids non-strict saddles, we develop trajectory-level control arguments for the end-to-end dynamics. In addition, we establish novel uniform concentration results that hold along the entire trajectory, and are essential for obtaining order-wise optimal sample complexity. We corroborate our theory with extensive experiments across a range of configurations.
Berk Tinaz, Changzhi Xie, Mahdi Soltanolkotabi
Department of Electrical and Computer Engineering University of Southern California Los Angeles, CA, USA · Department of Computer Science University of Southern California Los Angeles, CA, USA
We study the population loss landscape of two-layer ReLU networks of the form ∑k=1KReLU(wk⊤x) in a realisable teacher-student setting with Gaussian covariates. We show that local minima admit an exact low-dimensional representation in terms of summary statistics, yielding a sharp and interpretable characterisation of the landscape. We further establish a direct link with one-pass SGD: local minima correspond to attractive fixed points of the dynamics in summary statistics space. This perspective reveals a hierarchical organisation of minima into discrete families and shows how overparameterisation changes their stability and reachability under gradient-based dynamics. In this overparameterised regime, global minima become increasingly accessible, attracting the dynamics and reducing convergence to spurious solutions. Overall, our results reveal intrinsic limitations of common simplifying assumptions, which may miss essential features of the loss landscape even in minimal neural network models.
Jie Huang, Bruno Loureiro, Stefano Sarao Mannelli
Physics, Chalmers University of Technology and University of Gothenburg · Ecole Normale Superieure, PSL & CNRS · Data Science and AI, Computer Science and Engineering, Chalmers University of Technology and University of Gothenburg +1
Rich feature learning in tasks that unfold over time often requires the model to pass through bifurcations, constituting qualitative changes in the underlying model dynamics. We develop a local theory of gradient descent near these transitions through the empirical state-space neural tangent kernel (sNTK). Our central finding is that bifurcations both dominate and simplify learning dynamics: near bifurcations, we can reduce sNTK to a rank-one operator corresponding to learning in a classical normal form system, providing an analytically tractable description of the local learning geometry, even for high-dimensional recurrent systems. Concretely, we give a procedure for decomposing sNTK into bifurcation-relevant and residual channels, showing that near commonly codimension-1 bifurcations the relevant channel is a rank-one operator that is highly amplified. This amplification causes the bifurcation channel to dominate the full sNTK. Thus, bifurcations locally warp the learning landscape, funneling gradient descent into a few critical dynamical directions and making the nearby kernel and loss geometry predictable from classical normal forms. We illustrate this in a student-teacher recurrent neural network: the first learned bifurcation coincides with a sharp collapse in sNTK effective rank and the emergence of a dominant parameter direction whose restricted sNTK closely matches the landscape predicted by the scalar pitchfork normal form. Finally, we show that low-rank natural gradient methods resolve the resulting learning instability near bifurcations with very little overhead over SGD.