-
Competition-Level Code Generation with AlphaCode
Authors:
Yujia Li,
David Choi,
Junyoung Chung,
Nate Kushman,
Julian Schrittwieser,
Rémi Leblond,
Tom Eccles,
James Keeling,
Felix Gimeno,
Agustin Dal Lago,
Thomas Hubert,
Peter Choy,
Cyprien de Masson d'Autume,
Igor Babuschkin,
Xinyun Chen,
Po-Sen Huang,
Johannes Welbl,
Sven Gowal,
Alexey Cherepanov,
James Molloy,
Daniel J. Mankowitz,
Esme Sutherland Robson,
Pushmeet Kohli,
Nando de Freitas,
Koray Kavukcuoglu
, et al. (1 additional authors not shown)
Abstract:
Programming is a powerful and ubiquitous problem-solving tool. Develo** systems that can assist programmers or even generate programs independently could make programming more productive and accessible, yet so far incorporating innovations in AI has proven challenging. Recent large-scale language models have demonstrated an impressive ability to generate code, and are now able to complete simple…
▽ More
Programming is a powerful and ubiquitous problem-solving tool. Develo** systems that can assist programmers or even generate programs independently could make programming more productive and accessible, yet so far incorporating innovations in AI has proven challenging. Recent large-scale language models have demonstrated an impressive ability to generate code, and are now able to complete simple programming tasks. However, these models still perform poorly when evaluated on more complex, unseen problems that require problem-solving skills beyond simply translating instructions into code. For example, competitive programming problems which require an understanding of algorithms and complex natural language remain extremely challenging. To address this gap, we introduce AlphaCode, a system for code generation that can create novel solutions to these problems that require deeper reasoning. In simulated evaluations on recent programming competitions on the Codeforces platform, AlphaCode achieved on average a ranking of top 54.3% in competitions with more than 5,000 participants. We found that three key components were critical to achieve good and reliable performance: (1) an extensive and clean competitive programming dataset for training and evaluation, (2) large and efficient-to-sample transformer-based architectures, and (3) large-scale model sampling to explore the search space, followed by filtering based on program behavior to a small set of submissions.
△ Less
Submitted 8 February, 2022;
originally announced March 2022.
-
Data-driven model for the identification of the rock type at a drilling bit
Authors:
Nikita Klyuchnikov,
Alexey Zaytsev,
Arseniy Gruzdev,
Georgiy Ovchinnikov,
Ksenia Antipova,
Leyla Ismailova,
Ekaterina Muravleva,
Evgeny Burnaev,
Artyom Semenikhin,
Alexey Cherepanov,
Vitaliy Koryabkin,
Igor Simon,
Alexey Tsurgan,
Fedor Krasnov,
Dmitry Koroteev
Abstract:
Directional oil well drilling requires high precision of the wellbore positioning inside the productive area. However, due to specifics of engineering design, sensors that explicitly determine the type of the drilled rock are located farther than 15m from the drilling bit. As a result, the target area runaways can be detected only after this distance, which in turn, leads to a loss in well product…
▽ More
Directional oil well drilling requires high precision of the wellbore positioning inside the productive area. However, due to specifics of engineering design, sensors that explicitly determine the type of the drilled rock are located farther than 15m from the drilling bit. As a result, the target area runaways can be detected only after this distance, which in turn, leads to a loss in well productivity and the risk of the need for an expensive re-boring operation.
We present a novel approach for identifying rock type at the drilling bit based on machine learning classification methods and data mining on sensors readings. We compare various machine-learning algorithms, examine extra features coming from mathematical modeling of drilling mechanics, and show that the real-time rock type classification error can be reduced from 13.5 % to 9 %. The approach is applicable for precise directional drilling in relatively thin target intervals of complex shapes and generalizes appropriately to new wells that are different from the ones used for training the machine learning model.
△ Less
Submitted 25 March, 2019; v1 submitted 8 June, 2018;
originally announced June 2018.
-
τlepton decays and CVC
Authors:
V. A. Cherepanov,
S. I. Eidelman
Abstract:
We use experimental data on $e^+e^- \to η(η^{\prime})π^+π^-$ and conservation of vector current to estimate the branching fractions of $τ^-$ decay to $η(η^{\prime})π^-π^0ν_τ$. The obtained values are compared to the experimental results.
We use experimental data on $e^+e^- \to η(η^{\prime})π^+π^-$ and conservation of vector current to estimate the branching fractions of $τ^-$ decay to $η(η^{\prime})π^-π^0ν_τ$. The obtained values are compared to the experimental results.
△ Less
Submitted 21 April, 2009;
originally announced April 2009.
-
Critical temperature for the nuclear liquid-gas phase transition (from multifragmentation and fission)
Authors:
V. A. Karnaukhov,
H. Oeschler,
A. Budzanowski,
S. P. Avdeyev,
A. S. Botvina,
E. A. Cherepanov,
W. Karcz,
V. V. Kirakosyan,
P. A. Rukoyatkin,
I. Skwirczynska,
E. Norbeck
Abstract:
Critical temperature Tc for the nuclear liquid-gas phase transition is stimated both from the multifragmentation and fission data. In the first case,the critical temperature is obtained by analysis of the IMF yields in p(8.1 GeV)+Au collisions within the statistical model of multifragmentation (SMM). In the second case, the experimental fission probability for excited 188Os is compared with the…
▽ More
Critical temperature Tc for the nuclear liquid-gas phase transition is stimated both from the multifragmentation and fission data. In the first case,the critical temperature is obtained by analysis of the IMF yields in p(8.1 GeV)+Au collisions within the statistical model of multifragmentation (SMM). In the second case, the experimental fission probability for excited 188Os is compared with the calculated one with Tc as a free parameter. It is concluded for both cases that the critical temperature is higher than 16 MeV.
△ Less
Submitted 29 January, 2008;
originally announced January 2008.
-
Fission and Nuclear Liquid-Gas Phase Transition
Authors:
E. A. Cherepanov,
V. A. Karnaukhov
Abstract:
The temperature dependence of the liquid-drop fission barrier is considered, the critical temperature for the liquid-gas phase transition in nuclear matter being a parameter. Experimental and calculated data on the fission probability are compared for highly excited $^{188}$Os. The calculations have been made in the framework of the statistical model. It is concluded that the critical temperatur…
▽ More
The temperature dependence of the liquid-drop fission barrier is considered, the critical temperature for the liquid-gas phase transition in nuclear matter being a parameter. Experimental and calculated data on the fission probability are compared for highly excited $^{188}$Os. The calculations have been made in the framework of the statistical model. It is concluded that the critical temperature for the nuclear liquid--gas phase transition is higher than 16 MeV.
△ Less
Submitted 30 March, 2007;
originally announced March 2007.
-
Reaction 48Ca+208Pb: the capture-fission cross-sections and the mass-energy distributions of fragments above and deep below the Coulomb barrier
Authors:
E. V. Prokhorova,
E. A. Cherepanov,
M. G. Itkis,
N. A. Kondratiev,
E. M. Kozulin,
L. Krupa,
Yu. Ts. Oganessian,
V. V. Pashkevich,
I. V. Pokrovsky,
A. Ya. Rusanov
Abstract:
The capture-fission cross-sections in an energy range of 206-242 MeV of 48Ca-projectiles and mass-energy distributions (MEDs) of reaction products in an energy range of 211-242 MeV have been measured in the 48Ca+208Pb reaction using the double-arm time-of-flight spectrometer CORSET. The MEDs of fragments for heated fission were shown to consist of two components. One component, which is due to c…
▽ More
The capture-fission cross-sections in an energy range of 206-242 MeV of 48Ca-projectiles and mass-energy distributions (MEDs) of reaction products in an energy range of 211-242 MeV have been measured in the 48Ca+208Pb reaction using the double-arm time-of-flight spectrometer CORSET. The MEDs of fragments for heated fission were shown to consist of two components. One component, which is due to classical fusion-fission, is associated with the symmetric fission of the 256No compound nucleus. The other component, which appears as ''shoulders'', is associated with the quasi-fission process and can be named ''quasi-fission shoulders''. Those quasi-fission shoulders enclose light fragments whose masses are 60-90 a.m.u. The total kinetic energy (TKE) of the fragments that belong to the shoulders is higher than the value expected for a classical fusion-fission process. We have come to the conclusion that in quasi-fission, spherical shells with Z=28 and N=50 play a great role. It has also been demonstrated that the properties of the MEDs of fragments formally agree with a well-known hypothesis of two independent fission modes; in this case the modes are normal fusion-fission and quasi-fission processes. A high-energetic Super-Short mode of classical fission has been found at low excitation energies in the mass range of heavy fragments M = 130-135 and TKE = 233 MeV; however the yield associated with this mode is small.
△ Less
Submitted 29 September, 2003;
originally announced September 2003.
-
Dinuclear concept - cluster model of fusion
Authors:
G. G. Adamian,
N. V. Antonenko,
E. A. Cherepanov,
A. K. Nasirov,
W. Scheid,
V. V. Volkov
Abstract:
The synthesis of superheavy elements is analysed within the dinuclear system concept of compound nucleus formation. The perspectives for using radioactive beams in complete fusion reactions are discussed.
The synthesis of superheavy elements is analysed within the dinuclear system concept of compound nucleus formation. The perspectives for using radioactive beams in complete fusion reactions are discussed.
△ Less
Submitted 2 November, 1999;
originally announced November 1999.
