Carlos Riquelme Ruiz

11 papers A* 6Journal 5
YearRankTypeTitle / Venue / Authors
2024 A* conf
ICLR
Joan Puigcerver, Carlos Riquelme Ruiz, Basil Mustafa, Neil Houlsby
2024 A* conf
CVPR
Xi Chen, Josip Djolonga, Piotr Padlewski, Basil Mustafa, Soravit Changpinyo, Jialin Wu, Carlos Riquelme Ruiz, Sebastian Goodman, Xiao Wang, Yi Tay, Siamak Shakeri, Mostafa Dehghani, Daniel Salz, Mario Lucic, Michael Tschannen, Arsha Nagrani, Hexiang Hu, Mandar Joshi, Bo Pang, Ceslee Montgomery, Paulina Pietrzyk, Marvin Ritter, A. J. Piergiovanni, Matthias Minderer, Filip Pavetic, Austin Waters, Gang Li, Ibrahim Alabdulmohsin, Lucas Beyer, Julien Amelot, Kenton Lee, Andreas Peter Steiner, Yang Li, Daniel Keysers, Anurag Arnab, Yuanzhong Xu, Keran Rong, Alexander Kolesnikov, Mojtaba Seyedhosseini, Anelia Angelova, Xiaohua Zhai, Neil Houlsby, Radu Soricut
2024 J jnl
Trans. Mach. Learn. Res.
Tianlin Liu, Mathieu Blondel, Carlos Riquelme Ruiz, Joan Puigcerver
2024 A* conf
ICLR
Elias Frantar, Carlos Riquelme Ruiz, Neil Houlsby, Dan Alistarh, Utku Evci
2023 J jnl
CoRR
Xi Chen, Josip Djolonga, Piotr Padlewski, Basil Mustafa, Soravit Changpinyo, Jialin Wu, Carlos Riquelme Ruiz, Sebastian Goodman, Xiao Wang, Yi Tay, Siamak Shakeri, Mostafa Dehghani, Daniel Salz, Mario Lucic, Michael Tschannen, Arsha Nagrani, Hexiang Hu, Mandar Joshi, Bo Pang, Ceslee Montgomery, Paulina Pietrzyk, Marvin Ritter, A. J. Piergiovanni, Matthias Minderer, Filip Pavetic, Austin Waters, Gang Li, Ibrahim Alabdulmohsin, Lucas Beyer, Julien Amelot, Kenton Lee, Andreas Peter Steiner, Yang Li, Daniel Keysers, Anurag Arnab, Yuanzhong Xu, Keran Rong, Alexander Kolesnikov, Mojtaba Seyedhosseini, Anelia Angelova, Xiaohua Zhai, Neil Houlsby, Radu Soricut
2023 A* conf
ICML
Mostafa Dehghani, Josip Djolonga, Basil Mustafa, Piotr Padlewski, Jonathan Heek, Justin Gilmer, Andreas Peter Steiner, Mathilde Caron, Robert Geirhos, Ibrahim Alabdulmohsin, Rodolphe Jenatton, Lucas Beyer, Michael Tschannen, Anurag Arnab, Xiao Wang, Carlos Riquelme Ruiz, Matthias Minderer, Joan Puigcerver, Utku Evci, Manoj Kumar, Sjoerd van Steenkiste, Gamaleldin Fathy Elsayed, Aravindh Mahendran, Fisher Yu, Avital Oliver, Fantine Huot, Jasmijn Bastings, Mark Collier, Alexey A. Gritsenko, Vighnesh Birodkar, Cristina Nader Vasconcelos, Yi Tay, Thomas Mensink, Alexander Kolesnikov, Filip Pavetic, Dustin Tran, Thomas Kipf, Mario Lucic, Xiaohua Zhai, Daniel Keysers, Jeremiah J. Harmsen, Neil Houlsby
2023 A* conf
ICLR
Aran Komatsuzaki, Joan Puigcerver, James Lee-Thorp, Carlos Riquelme Ruiz, Basil Mustafa, Joshua Ainslie, Yi Tay, Mostafa Dehghani, Neil Houlsby
2022 J jnl
Trans. Mach. Learn. Res.
James Urquhart Allingham, Florian Wenzel, Zelda E. Mariet, Basil Mustafa, Joan Puigcerver, Neil Houlsby, Ghassen Jerfel, Vincent Fortuin, Balaji Lakshminarayanan, Jasper Snoek, Dustin Tran, Carlos Riquelme Ruiz, Rodolphe Jenatton
2022 J jnl
CoRR
Aran Komatsuzaki, Joan Puigcerver, James Lee-Thorp, Carlos Riquelme Ruiz, Basil Mustafa, Joshua Ainslie, Yi Tay, Mostafa Dehghani, Neil Houlsby
2021 A* conf
ICLR
Joan Puigcerver, Carlos Riquelme Ruiz, Basil Mustafa, Cédric Renggli, André Susano Pinto, Sylvain Gelly, Daniel Keysers, Neil Houlsby
2021 J jnl
CoRR
James Urquhart Allingham, Florian Wenzel, Zelda E. Mariet, Basil Mustafa, Joan Puigcerver, Neil Houlsby, Ghassen Jerfel, Vincent Fortuin, Balaji Lakshminarayanan, Jasper Snoek, Dustin Tran, Carlos Riquelme Ruiz, Rodolphe Jenatton
redb/extractors/elf_extractors/elf_sections.py
← Index redb/extractors/elf_extractors/elf_sections.py python
import inspect
import hashlib
import math
from collections import Counter
from datetime import datetime, timezone
from typing import Any, List, Dict

from elftools.elf.elffile import ELFFile
from elftools.common.exceptions import ELFError

from redb.extractors.enum import Tag
from redb.extractors.elf_extractor import ELFExtractor
from redb.models.dataclasses import ELFSection


class ELFSectionExtractor(ELFExtractor):

    def __init__(
        self,
        filepath,
        log,
        exporters=None,
        index_prefix=None,
        elastic_index=None,
        known_benign=False,
        known_malicious=False,
        elf=None,
    ):
        super().__init__(
            filepath,
            log,
            exporters,
            index_prefix,
            elastic_index,
            known_benign,
            known_malicious,
            elf,
        )
        self.elf_sections = []
        self.elastic_index = self.index_prefix + "-elf_sections"
        self.log.debug(inspect.currentframe().f_code.co_name)

    def _is_empty_result(self, extracted_data) -> bool:
        """
        Override: Empty sections is an ERROR, not a valid empty case.
        A valid ELF file must have sections (at minimum a null section).
        """
        # Always return False - empty sections should be treated as an error
        return False

    def _calculate_entropy(self, data: bytes) -> float:
        """Calculate Shannon entropy of data."""
        if not data:
            return 0.0

        try:
            # Count frequency of each byte
            byte_counts = Counter(data)
            data_len = len(data)

            # Calculate entropy
            entropy = 0.0
            for count in byte_counts.values():
                if count > 0:
                    frequency = count / data_len
                    entropy -= frequency * math.log2(frequency)

            return entropy
        except Exception as e:
            self.log.error(f"Error calculating entropy: {e}")
            return 0.0

    def _map_section_type(self, sh_type_str: str) -> int:
        """Map section type string to enum value."""
        type_map = {
            'SHT_NULL': 0,
            'SHT_PROGBITS': 1,
            'SHT_SYMTAB': 2,
            'SHT_STRTAB': 3,
            'SHT_RELA': 4,
            'SHT_HASH': 5,
            'SHT_DYNAMIC': 6,
            'SHT_NOTE': 7,
            'SHT_NOBITS': 8,
            'SHT_REL': 9,
            'SHT_DYNSYM': 11
        }
        return type_map.get(sh_type_str, 0)

    def _decode_section_flags(self, flags: int) -> List[str]:
        """Decode section flags to human-readable strings."""
        flag_strings = []

        # Common ELF section flags
        if flags & 0x1:  # SHF_WRITE
            flag_strings.append('WRITE')
        if flags & 0x2:  # SHF_ALLOC
            flag_strings.append('ALLOC')
        if flags & 0x4:  # SHF_EXECINSTR
            flag_strings.append('EXECINSTR')
        if flags & 0x10:  # SHF_MERGE
            flag_strings.append('MERGE')
        if flags & 0x20:  # SHF_STRINGS
            flag_strings.append('STRINGS')
        if flags & 0x40:  # SHF_INFO_LINK
            flag_strings.append('INFO_LINK')
        if flags & 0x80:  # SHF_LINK_ORDER
            flag_strings.append('LINK_ORDER')
        if flags & 0x100:  # SHF_OS_NONCONFORMING
            flag_strings.append('OS_NONCONFORMING')
        if flags & 0x200:  # SHF_GROUP
            flag_strings.append('GROUP')
        if flags & 0x400:  # SHF_TLS
            flag_strings.append('TLS')

        return flag_strings if flag_strings else ['NONE']

    def _extract_section_data(self, section) -> Dict:
        """Extract data from a single section."""
        try:
            header = section.header

            # Get section name (handle empty names)
            section_name = section.name if section.name else f"<unnamed_{section.header.get('sh_name', 0)}>"

            # Get section type and map to enum
            sh_type_str = header.get('sh_type', 'SHT_NULL')
            section_type_enum = self._map_section_type(sh_type_str)
            section_type_str = sh_type_str.replace('SHT_', '') if sh_type_str.startswith('SHT_') else sh_type_str

            # Get section properties
            section_flags = header.get('sh_flags', 0)
            section_flags_str = self._decode_section_flags(section_flags)
            section_addr = header.get('sh_addr', 0)
            section_offset = header.get('sh_offset', 0)
            section_size = header.get('sh_size', 0)
            section_link = header.get('sh_link', 0)
            section_info = header.get('sh_info', 0)
            section_addralign = header.get('sh_addralign', 0)
            section_entsize = header.get('sh_entsize', 0)

            # Calculate entropy and hashes for section data
            section_entropy = 0.0
            section_sha256 = ""
            section_md5 = ""

            try:
                if section_size > 0 and section_type_str != 'NOBITS':
                    section_data = section.data()
                    if section_data:
                        # Calculate entropy
                        section_entropy = self._calculate_entropy(section_data)

                        # Calculate hashes
                        section_sha256 = hashlib.sha256(section_data).hexdigest()
                        section_md5 = hashlib.md5(section_data).hexdigest()
            except Exception as e:
                self.log.warning(f"Could not read section '{section_name}' data: {e}")
                section_entropy = 0.0
                section_sha256 = ""
                section_md5 = ""

            return ELFSection(
                section_name=section_name,
                section_type=section_type_enum,
                section_type_str=section_type_str,
                section_flags=section_flags,
                section_flags_str=section_flags_str,
                section_addr=section_addr,
                section_offset=section_offset,
                section_size=section_size,
                section_link=section_link,
                section_info=section_info,
                section_addralign=section_addralign,
                section_entsize=section_entsize,
                section_entropy=section_entropy,
                section_sha256=section_sha256,
                section_md5=section_md5
            )

        except Exception as e:
            self.log.error(f"Error extracting section data: {e}")
            return None

    def tag(self):
        return Tag.ELF_SECTIONS.value if hasattr(Tag, 'ELF_SECTIONS') else "elf_sections"

    def extract(self):
        try:
            self.log.debug(inspect.currentframe().f_code.co_name)

            def extract_data(elf):
                sections_data = []

                # Iterate through all sections with per-section error handling
                for section_index, section in enumerate(elf.iter_sections()):
                    try:
                        section_data = self._extract_section_data(section)
                        if section_data:
                            sections_data.append(section_data)
                        else:
                            self.log.warning(f"Failed to extract data for section {section_index}")
                    except Exception as e:
                        self.log.warning(f"Error processing section {section_index}: {e}")
                        # Continue processing other sections

                return sections_data

            if not self._is_elf_file():
                return None

            result = self._with_elf_file(extract_data)
            if result is None:
                return None

            self.elf_sections = result
            return self.elf_sections

        except Exception as e:
            self.log.error(f"Error extracting ELF sections {self.hash.sha256}: {e}")
            return None

    def prepare_export_data(self, exporter_type: str) -> Any:
        self.log.debug(inspect.currentframe().f_code.co_name)

        if exporter_type == "ElasticsearchExporter":
            return self.elf_sections
        elif exporter_type == "ClickHouseExporter":
            try:
                if not self.elf_sections:
                    return None

                # Prepare data arrays for all sections
                data = []
                current_time = datetime.now(timezone.utc)
                for section in self.elf_sections:
                    row = [
                        self.sha256,
                        self.md5,
                        self.sha1,
                        section.section_name,
                        section.section_type,
                        section.section_type_str,
                        section.section_flags,
                        section.section_flags_str,
                        section.section_addr,
                        section.section_offset,
                        section.section_size,
                        section.section_link,
                        section.section_info,
                        section.section_addralign,
                        section.section_entsize,
                        section.section_entropy,
                        section.section_sha256,
                        section.section_md5,
                        current_time
                    ]
                    data.append(row)

                column_names = [
                    'sha256', 'md5', 'sha1',
                    'section_name', 'section_type', 'section_type_str',
                    'section_flags', 'section_flags_str',
                    'section_addr', 'section_offset', 'section_size',
                    'section_link', 'section_info', 'section_addralign', 'section_entsize',
                    'section_entropy', 'section_sha256', 'section_md5',
                    'analysis_date'
                ]

                column_type_names = [
                    'FixedString(64)', 'FixedString(32)', 'FixedString(40)',
                    'LowCardinality(String)',
                    "Enum8('NULL'=0, 'PROGBITS'=1, 'SYMTAB'=2, 'STRTAB'=3, 'RELA'=4, 'HASH'=5, 'DYNAMIC'=6, 'NOTE'=7, 'NOBITS'=8, 'REL'=9, 'DYNSYM'=11)",
                    'LowCardinality(String)',
                    'UInt64',
                    'Array(LowCardinality(String))',
                    'UInt64', 'UInt64', 'UInt64',
                    'UInt32', 'UInt32', 'UInt64', 'UInt64',
                    'Float64',
                    'FixedString(64)', 'FixedString(32)',
                    'DateTime64(3, \'UTC\')'
                ]

                if not data:
                    return None

                return (data, column_names, column_type_names)

            except Exception as e:
                self.log.error(f"Error preparing export data: {e}")
                raise

    def get_clickhouse_table(self) -> str:
        return "redb_elf_sections"