Deep learning reconstruction of neutrino direction, energy, and flavor with complete uncertainty predictions

Nils Heyer,Thorsten Glüsenkamp,Christian Glaser

2024-10-12

Abstract:With the IceCube-Gen2 observatory under development and RNO-G under construction, the first detection of ultra-high-energy neutrinos is on the horizon making event reconstruction a priority. Here, we present a full reconstruction of the neutrino direction, shower energy, and interaction type (and thereby flavor) from raw antenna signals. We use a deep neural network with conditional normalizing-flows for the reconstruction. This, for the first time, allows for event-by-event predictions of the posterior distribution of all reconstructed properties, in particular, the asymmetric uncertainties of the neutrino direction. The algorithm was applied to an extensive MC dataset of 'shallow' and 'deep' detector components in South Pole ice. We present the reconstruction performance and compare the two station components. For the first time, we quantify the effect of birefringence on event reconstruction.

Instrumentation and Methods for Astrophysics,High Energy Astrophysical Phenomena

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: how to use deep - learning techniques to accurately reconstruct the direction, energy, and flavor (i.e., interaction type) of ultra - high - energy neutrinos, and provide complete uncertainty predictions. Specifically, the key points of the paper include:

1. **Detection and Reconstruction of Ultra - High - Energy Neutrinos**:
- With the development of the IceCube - Gen2 observatory and the construction of RNO - G, the possibility of detecting ultra - high - energy neutrinos for the first time is gradually increasing, so event reconstruction has become a priority task.
- The author proposes a method using deep neural networks (especially conditional normalization flows) to reconstruct the direction, energy, and interaction type of neutrinos (and thus infer their flavor) from the original antenna signals.

2. **Uncertainty Prediction**:
- Traditional deep - learning methods usually can only provide the best - fit values and are difficult to quantify uncertainties. The method proposed in this paper can predict the posterior distribution for each event, especially the asymmetric uncertainty of the neutrino direction.
- This is the first time to achieve the prediction of the posterior distribution for all reconstructed properties on an event - by - event basis.

3. **Impact of Birefringence Effects**:
- The Antarctic ice sheet has birefringent properties, which will affect the propagation of radio pulses. The author has studied the impact of birefringence effects on event reconstruction and retrained a model to correct these effects.
- The results show that although the accuracy of direction reconstruction has slightly decreased, the accuracy of energy reconstruction has significantly increased.

4. **Performance Comparison of Different Detector Components**:
- The performance of two differently designed detector components (shallow - layer and deep - layer) in neutrino reconstruction has been studied, and a detailed performance comparison has been carried out.

In summary, this paper aims to show the potential of current and future sub - ice radio neutrino detectors in reconstruction, especially in the reconstruction of ultra - high - energy neutrino events, how to improve the accuracy and quantify the uncertainties through deep - learning techniques.