SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2211.14939 · NeurIPS · 2022

Applying Deep Reinforcement Learning to the HP Model for Protein Structure Prediction

Kaiyuan Yang, Houjing Huang, Olafs Vandans, Adithya Murali, Fujia Tian, Roland Yap, Liang Dai

arXiv · PDF · Open in the Atlas

Code that ran

11 of 13 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 11 have no contract verdict; 0 failures of our instrument; 2 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2211.14939.json · badge: /api/ran/2211.14939.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction canonical (classifier label)
classifier basis not recorded on the link
11 ran · 2 no recorded run
FunctionStatusWhere it lives
CosineAnnealing Ran
ran on a synthesized input
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/code/annealing_schedule.py
code served (permissive licence) · get_code("1cf642d8e1180240")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import math

def CosineAnnealing(epoch, T_input, eta_min=0.01, eta_max=1.0, warmRestart=True):
    """
    T_input for T_{0,i,cur,max} etc
    warmRestart: whether to have automatic cyclic warmRestart
    """
    if warmRestart:
        T_0 = T_input
        T_i = T_0
        if epoch >= T_0:
            T_cur = epoch % T_0
        else:
            T_i = T_0
            T_cur = epoch
        eta_t = eta_min + (eta_max - eta_min) * (1 + math.cos(math.pi * T_cur / T_i)) / 2
    else:
        T_max = T_input
        T_cur = epoch
        eta_t = eta_min + (eta_max - eta_min) * (1 + math.cos(math.pi * T_cur / T_max)) / 2
    return eta_t


ARGS = {
    "epoch": 5,
    "T_input": 10,
    "eta_min": 0.01,
    "eta_max": 1.0,
    "warmRestart": True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
CosineAnnealingSchedule Ran
ran on a synthesized input
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/code/annealing_schedule.py
code served (permissive licence) · get_code("9b255e5def761d98")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import math

# CosineAnnealingSchedule parameters
ARGS = {
    "epoch": 5,
    "num_episodes": 100,
    "eta_min": 0.01,
    "eta_max": 1.0,
    "warmRestart": True,
    "num_restarts": 10
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
avg_std_of_max Ran
ran on a synthesized input
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/code/plot_rewards.py
code served (permissive licence) · get_code("4a48db84d98db28a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Analyze the function to understand what parameters are needed:
# - seed_42_data, seed_1984_data, seed_1991_data, seed_2021_data: numpy arrays of rewards
# - num_episodes: integer, total number of episodes
# - N_chunks: integer, number of chunks (derived from num_episodes // show_every)
# - show_every: integer, episodes per chunk

# The function splits rewards into num_episodes // show_every chunks
# So we need: num_episodes to be divisible by show_every
# And N_chunks should equal num_episodes // show_every

num_episodes = 100
show_every = 10
N_chunks = num_episodes // show_every  # = 10

# Create reward arrays with shape (num_episodes,)
# Each element represents a reward for that episode
seed_42_data = np.random.RandomState(42).uniform(0, 100, num_episodes)
seed_1984_data = np.random.RandomState(1984).uniform(0, 100, num_episodes)
seed_1991_data = np.random.RandomState(1991).uniform(0, 100, num_episodes)
seed_2021_data = np.random.RandomState(2021).uniform(0, 100, num_episodes)

ARGS = {
    'seed_42_data': seed_42_data,
    'seed_1984_data': seed_1984_data,
    'seed_1991_data': seed_1991_data,
    'seed_2021_data': seed_2021_data,
    'num_episodes': num_episodes,
    'N_chunks': N_chunks,
    'show_every': show_every,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
delta_ESB_OS Ran
ran on a synthesized input
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/Waseda_Gym_Lattice/early_stop.py
code served (permissive licence) · get_code("cfe96fadd0fed716")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Analyzing the function:
# - E_S_B: numeric value used in comparisons and arithmetic
# - O_S: numeric value used in comparisons and arithmetic
# - E_terminal_H: boolean value checked with 'and' operator
# - O_terminal_H: boolean value checked with 'and' operator

ARGS = {
    "E_S_B": 5,
    "O_S": 3,
    "E_terminal_H": True,
    "O_terminal_H": False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
delta_OSB_ES Ran
ran on a synthesized input
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/Waseda_Gym_Lattice/early_stop.py
code served (permissive licence) · get_code("38d7a166b731d649")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Analyzing the function delta_OSB_ES:
# - Takes 4 parameters: O_S_B, E_S, O_terminal_H, E_terminal_H
# - O_S_B and E_S are used in arithmetic operations (*, +, -, min, max, ==, >=)
#   so they should be numeric (int or float)
# - O_terminal_H and E_terminal_H are used in boolean context (if E_terminal_H, if O_terminal_H)
#   so they should be booleans

ARGS = {
    "O_S_B": 5,
    "E_S": 3,
    "O_terminal_H": True,
    "E_terminal_H": False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
derive_LFR_direction Ran
ran on a synthesized input
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/code/hpsandbox_util.py
code served (permissive licence) · get_code("e7a0a41e3d3cc24a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

ARGS = {
    'p1': (0, 0),
    'p2': (1, 0),
    'p3': (1, 1)
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_max_per_chunk Ran
ran on a synthesized input
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/code/plot_rewards.py
code served (permissive licence) · get_code("16c61ce5e9555d27")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (4,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct arguments for extract_max_per_chunk
# The function expects:
# - rewards_all_episodes: a 1D numpy array of rewards
# - num_episodes: an integer number of episodes
# - show_every: an integer for chunking

num_episodes = 12
show_every = 3

# Create a 1D array of rewards with num_episodes elements
# Using distinct values so we can verify max extraction works
rewards_all_episodes = np.array([1.5, 2.3, 0.8, 3.1, 1.2, 2.9, 0.5, 4.2, 1.8, 2.1, 3.5, 0.9])

ARGS = {
    "rewards_all_episodes": rewards_all_episodes,
    "num_episodes": num_episodes,
    "show_every": show_every
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (4,) — shape (4,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
label_conf_seq Ran
ran on a synthesized input
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/code/hpsandbox_util.py
code served (permissive licence) · get_code("fc875ade9a7a3423")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

ARGS = {
    "conf_seq": [(0, 0), (0, 1), (3, 1), (2, 2)],
    "conf_title": "HHPPH.txt"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list — a list output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
move_LFR_direction Ran
ran on a synthesized input
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/code/hpsandbox_util.py
code served (permissive licence) · get_code("57d5b400f5405c0b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family int; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Test case 1: moving forward
# p1 = (0, 0), p2 = (1, 0), move forward means next point is (2, 0)
# Vector from p1 to p2 is (1, 0)
# Moving forward means same direction, so p3 should be (2, 0)
p1_test = (0, 0)
p2_test = (1, 0)
move_direction_test = 1  # forward

ARGS = {
    'p1': p1_test,
    'p2': p2_test,
    'move_direction': move_direction_test
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family int — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
secondsToStr Ran
ran on a synthesized input
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/Waseda_Gym_Lattice/timer.py
code served (permissive licence) · get_code("95240391ed42ab43")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

from time import strftime, localtime
from datetime import timedelta

# The function secondsToStr takes one optional parameter: elapsed
# If elapsed is None, it returns the current date-time string
# If elapsed is a number, it converts seconds to a timedelta string

# Test case 1: with elapsed=None (default)
ARGS = {
    "elapsed": None
}

# Alternative test case: with elapsed as a positive number
# ARGS = {
#     "elapsed": 3661  # 1 hour, 1 minute, 1 second
# }

# Using the first version which tests the None case
ARGS = {
    "elapsed": None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
seq_parity_stats Ran
ran on a synthesized input
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/Waseda_Gym_Lattice/early_stop.py
code served (permissive licence) · get_code("b8267191de192c1f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

def seq_parity_stats(seq):
    Odd_H_indices = []
    Even_H_indices = []
    O_S, E_S = (0, 0)
    O_terminal_H = False
    E_terminal_H = False
    for index, aa in enumerate(seq):
        order = index + 1
        print(f'Order-{order}-Residue-{aa}')
        if aa == 'H':
            if order % 2 == 1:
                print('H Odd')
                Odd_H_indices.append(order)
                O_S += 1
                if order == 1 or order == len(seq):
                    O_terminal_H = True
            elif order % 2 == 0:
                print('H Even')
                Even_H_indices.append(order)
                E_S += 1
                if order == len(seq):
                    E_terminal_H = True
    Odd_H_indices = np.asarray(Odd_H_indices, dtype=int)
    Even_H_indices = np.asarray(Even_H_indices, dtype=int)
    print('Odd_H_indices = ', Odd_H_indices)
    print('Even_H_indices = ', Even_H_indices)
    print('O_S = ', O_S)
    print('E_S = ', E_S)
    print('O_terminal_H = ', O_terminal_H)
    print('E_terminal_H = ', E_terminal_H)
    OPT_S = 2 * min(O_S, E_S) + 2
    print('OPT_S = ', OPT_S)
    return (Odd_H_indices, Even_H_indices, O_S, E_S, O_terminal_H, E_terminal_H, OPT_S)

# Construct valid arguments for seq_parity_stats
# The function expects a sequence (seq) which it iterates over
# It checks for 'H' characters at odd and even positions
# A simple string or list of characters works

seq_example = "HXHXHXHX"

ARGS = {
    "seq": seq_example
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
ExponentialDecay Not yet run
unverified (harvested, no recorded run)
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/code/annealing_schedule.py
code served (permissive licence) · get_code("87cdf5b45663387d")
How it ran

no invocation recorded.

plot_moving_avg Not yet run
unverified (harvested, no recorded run)
compsoftmatterbiophysics-cityu-hk/applying-drl-to-hp-model-for-protein-structure-prediction/code/plot_rewards.py
code served (permissive licence) · get_code("676ae37285091875")
How it ran

no invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 13 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

A central problem in computational biophysics is protein structure prediction, i.e., finding the optimal folding of a given amino acid sequence. This problem has been studied in a classical abstract model, the HP model, where the protein is modeled as a sequence of H (hydrophobic) and P (polar) amino acids on a lattice. The objective is to find conformations maximizing H-H contacts. It is known that even in this reduced setting, the problem is intractable (NP-hard). In this work, we apply deep reinforcement learning (DRL) to the two-dimensional HP model. We can obtain the conformations of best known energies for benchmark HP sequences with lengths from 20 to 50. Our DRL is based on a deep Q-network (DQN). We find that a DQN based on long short-term memory (LSTM) architecture greatly enhances the RL learning ability and significantly improves the search process. DRL can sample the state space efficiently, without the need of manual heuristics. Experimentally we show that it can find multiple distinct best-known solutions per trial. This study demonstrates the effectiveness of deep reinforcement learning in the HP model for protein folding.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2211.14939")
get_code_for_paper("2211.14939")
have("2211.14939")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2211.14939.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2211.14939.svg)](https://syntology.ai/paper/2211.14939)

Connect an agent — have() is free.