Papers for

machine learning hardware developers

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Backward state choice affects learning outcomes in low precision training

Backward-State Policy Is Part of the Learning Algorithm

Abstract: Low-precision training rounds tensors that the backward pass reads again, often for several gradients; each use can read the forward's rounded value, the original, or a new random rounding. This backward-state policy looks like a memory and precision detail, settled by copy accuracy and final loss. We argue that it is part of the learning algorithm, and that neither check shows whether it is right. Copy accuracy does not decide the outcome: in three pairs of 390M runs with an emulated FP8 backward, training fails when attention's backward reuses the forward's rounded output and succeeds with a new rounding from the same distribution. Even the most accurate copy, the original itself, can be wrong by our reference: the gradient of the forward pass as it actually ran, with gradients passed through rounding unchanged. For example, a normalization output stored in low precision feeds two gradients: the gain's gradient needs the original, but the next layer's weight gradient needs the rounded value that layer multiplied. Final loss, the other check, does not rule out the error of reading the original for both: it persists in models trained with such a store, while planned loss comparisons stay within a margin fixed in advance. We therefore derive from this reference which value each use must read, or which substitute gives the same gradient on average with the forward held fixed, and check these per-use requirements on single operators, without training. In three tests using PyTorch and Transformer Engine, the requirements predicted beforehand whether reuse changes what the backward computes on average relative to an independent copy, and every prediction held. Backward-state policy is thus part of the learning algorithm: it should be specified and checked use by use, not settled by copy accuracy and final loss.

Wed 30 SeptMachine Learning
The gist
Training AI models often uses shortcuts when copying numbers during calculations, especially in lower-precision formats to save space and time. The authors found that how these copied numbers are reused in backpropagation, specifically whether the backward step uses the same rounded number or a fresh one, actually changes the learning result. This means the choice of which value to use in each step is not just a technical detail but part of how the model learns. They proposed a way to decide which value to use for each calculation, making these choices explicit and testable without full retraining.
Open → 2609.39813v1

Analog dnn inference accuracy improves with temperature aware methods

Beyond Noise: Understanding and Overcoming Temperature Effects in Analog DNN Inference

Abstract: The energy efficiency of analog computing makes it one of the most promising candidates for deploying resource-intensive machine learning workloads on constrained platforms such as mobile and embedded devices. However, analog accelerators are inherently susceptible to noise and non-idealities arising from physical component variations, whose behavior is further sensitive to environmental factors. These effects can significantly degrade inference accuracy. In this work, we conduct a comprehensive experimental study on a representative example of analog hardware to investigate the impact of temperature. We first characterize the behavior of stochastic and systematic non-idealities across a range of operating temperatures. Following this, we compare a set of simulation-based and hardware-based mitigation strategies aimed at improving robustness against temperature-induced performance degradation. Our results suggest that temperature-induced degradation is driven primarily by systematic non-idealities rather than stochastic noise alone. Noise-aware training improves robustness, while hardware-in-the-loop training and temperature-aware calibration provide the strongest accuracy retention across varying thermal conditions.

Mon 14 SeptMachine Learning
The gist
Analog computing devices can use less energy than regular digital ones, making them good for phones and embedded gadgets. But they can get less accurate when the temperature changes because parts behave differently. The authors studied how temperature affects these devices and found that consistent changes in components cause most problems, more than random noise. They tested different training and calibration methods and found that training with hardware and accounting for temperature help keep accuracy steady.
Open → 2609.15527v1