UniSymNet:疎なエンコーディングと二層最適化を備えた統一シンボリックネットワーク
Xinxin Li1, Juan Zhang2, Da Li3
1School of Mathematical Sciences, East China Normal University, Shang Hai, 200241, China.
まとめ
この研究では、数学的表現の発見を改善するための統一シンボリックネットワーク(UniSymNet)を紹介します。UniSymNetは、複雑性を低減し、自然現象のモデリングにおけるパフォーマンスを向上させるために演算子を統一します。
科学分野:
- 人工知能
- 機械学習
- 計算数学
背景:
- シンボリック回帰(SR)は、自然現象をモデル化するための数学的表現を発見する上で重要です。
- 従来のSR方法は、複雑なツリー構造に苦労し、パフォーマンスを制限します。
- 既存のシンボリックネットワークは、多変量演算子と固定アーキテクチャの課題に直面しており、過剰適合につながります。
研究 の 目的:
- 既存のシンボリックネットワークの限界を克服する新しい統一シンボリックネットワーク(UniSymNet)を提案すること。
- 発見された数学的表現の表現力を高め、複雑性を低減すること。
- 多様なデータセットに適応可能な柔軟なトレーニングフレームワークを開発すること。
主な方法:
- 非線形二項演算子をネストされた単項演算子に統合して、多変量演算子を作成します。
- 二層最適化フレームワークの開発:構造選択のためのTransformer事前トレーニングと目的固有のパラメータ最適化。
- UniSymNetの機能確立のための厳密な理論的証明。
主要な成果:
- UniSymNetは、複雑性と表現力において理論的な利点を示します。
- 二層最適化フレームワークにより、柔軟な構造適応と式の複雑性の低減が可能になります。
- 標準ベンチマークとSRBenchでの評価により、高いシンボリック解率、適合精度、および低い複雑性が示されます。
結論:
- UniSymNetは、強化された機能を持つシンボリック回帰のための有望な新しいパラダイムを提供します。
- 提案されたトレーニングフレームワークは、構造選択とパラメータ最適化を効果的にガイドします。
- UniSymNetは、正確で簡潔な数学的表現を発見する上で優れたパフォーマンスを達成します。
関連する概念動画
Lewis Symbols and the Octet Rule
80.9K
Chemical bonds are complex interactions between two or more atoms or ions, which reduce the potential energy of the molecule. Gilbert N. Lewis developed a model called the Lewis model that simplified the depiction of chemical bond formation and provided straightforward explanations for the chemical bonds seen in most common compounds.
80.9K
Elements: Chemical Symbols and Isotopes
126.1K
A chemical symbol is an abbreviation used to indicate an element or an atom of an element. For example, the symbol for mercury is Hg. The same symbol is used to indicate one atom of mercury (microscopic domain) or to label a container of many atoms of the element mercury (macroscopic domain).
Some symbols are derived from the common English name of the element; others are abbreviations of the name in another language — Latin, Greek or German. For example, the symbol for aluminum (common name)...
Some symbols are derived from the common English name of the element; others are abbreviations of the name in another language — Latin, Greek or German. For example, the symbol for aluminum (common name)...
126.1K
Chemical Symbols
53.3K
A chemical symbol is an abbreviation that is used to indicate an element or an atom of an element. For example, the symbol for mercury is Hg. We use the same symbol to indicate one atom of mercury (microscopic domain) or to label a container of many atoms of the element mercury (macroscopic domain).
Some symbols are derived from the common name of the element; others are abbreviations of the name in another language. Most symbols have one or two letters, but three-letter symbols have been used...
Some symbols are derived from the common name of the element; others are abbreviations of the name in another language. Most symbols have one or two letters, but three-letter symbols have been used...
53.3K
Encoding
850
Information enters the brain through encoding, which is the input of information into the memory system. Once sensory information is received from the environment, the brain labels or codes it. The information is then organized with similar information and connected to existing concepts. Encoding occurs through automatic processing and effortful processing.
Automatic processing involves the encoding of details like time, space, frequency, and the meaning of words, usually done without conscious...
Automatic processing involves the encoding of details like time, space, frequency, and the meaning of words, usually done without conscious...
850
Protein Networks
4.5K
An organism can have thousands of different proteins, and these proteins must cooperate to ensure the health of an organism. Proteins bind to other proteins and form complexes to carry out their functions. Many proteins interact with multiple other proteins creating a complex network of protein interactions.
These interactions can be represented through maps depicting protein-protein interaction networks, represented as nodes and edges. Nodes are circles that are representative of a protein,...
These interactions can be represented through maps depicting protein-protein interaction networks, represented as nodes and edges. Nodes are circles that are representative of a protein,...
4.5K
Network Covalent Solids
16.2K
Network covalent solids contain a three-dimensional network of covalently bonded atoms as found in the crystal structures of nonmetals like diamond, graphite, silicon, and some covalent compounds, such as silicon dioxide (sand) and silicon carbide (carborundum, the abrasive on sandpaper). Many minerals have networks of covalent bonds.
To break or to melt a covalent network solid, covalent bonds must be broken. Because covalent bonds are relatively strong, covalent network solids are typically...
To break or to melt a covalent network solid, covalent bonds must be broken. Because covalent bonds are relatively strong, covalent network solids are typically...
16.2K


